Pesticide residues in food - 2003 - Joint FAO/WHO Meeting on Pesticide Residues

First draft prepared by
T.C. Marrs & A. Adjei
Food Standards Agency, London, England

Paraquat is a bipyridilium herbicide that was evaluated by the JMPR in 1970, 1972, 1976, 1985 and 1986 (Annex 1, references 14, 18, 26, 47), in order to establish an acceptable daily intake (ADI). A toxicological monograph was published after the 1970 JMPR and addenda to the monograph were published after the 1972, 1976 and 1982 Meetings. A toxicological monograph was published after the 1986 JMPR. At the JMPR in 1970, an ADI of 0-0.001 mg/kg bw, as paraquat dichloride, was established. The 1972 JMPR assigned an ADI of 0-0.002 mg/kg bw, while the 1982 JMPR reduced the ADI to 0-0.001 mg/kg bw. The 1986 JMPR established an ADI of 0-0.004 mg/kg bw as paraquat ion (equal to 0-0.006 mg/kg bw as the dichloride).

Paraquat was re-evaluated by the present Meeting within the periodic review programme of the Codex Committee on Pesticide Residues. A considerable amount of data has been generated since 1986 and was submitted for evaluation; these data include studies on the absorption, distribution, metabolism and excretion of paraquat and numerous studies of toxicity (acute, reproductive and developmental). Furthermore, a substantial number of papers in the open literature on, inter alia, the genotoxicity and neurotoxicity of paraquat have been reviewed. In all studies relevant to risk assessment, doses and intakes are expressed as paraquat ion.¹

1.1 Absorption, distribution and excretion

Rats

In a study of the absorption, distribution and excretion of paraquat, a single oral dose of ¹⁴C-labelled paraquat ion at 1 mg/kg bw was administered to five male and five female Alpk: ApfSD rats by gavage. Paraquat dichloride was used as the test material; the purity of the ¹⁴C-labelled material was 100%, while that of the unlabelled material was >96%. The specific activity of the radiolabelled material was 4.0996 GBq/mmol and that of the dosing solution was 4.12 MBq/g. Urine was collected 6 h after dosing and urine and faeces were collected separately at 12, 24, 36, 48 and 72 h after dosing. The animals were killed after 3 days and selected organs and tissues were removed. The amount of radioactivity remaining in the blood, selected tissues and the carcasses was estimated. Excretion of the radiolabel was rapid: in the first 24 h, in males 17.9% of the dose was excreted in the urine and 63.1% in the faeces. Equivalent figures for females were 11.6% and 74.1%. More than 90% of the radiolabel was eliminated in 72 h in both sexes. More radiolabel was excreted in the faeces of females than males. Only low concentrations of radiolabel were retained in the residual carcasses (0.64% and 0.54% of the administered dose in male and females respectively), the highest concentrations (0.01-0.02%) being found in the liver, lungs and kidneys (Lythgoe & Howard, 1995a).

In a second study of the absorption, distribution and excretion of paraquat, daily oral doses of paraquat (1 mg of paraquat ion/kg bw) were administered by gavage to eight male and eight female Alpk: ApfSD rats for 14 days. Paraquat dichloride was used as the test material; the purity of the ¹⁴C-labelled material was 100%, while that of the unlabelled material was >96%. A single oral dose of ¹⁴C-labelled paraquat ion at 1 mg/kg bw was subsequently administered by gavage. The specific activity of the radiolabelled material was 4.0996 GBq/mmol and that of the dosing solution was 4.12 MBq/g. Urine was collected 6 h after dosing and urine and faeces were collected separately at 12, 24, 36, 48 and 72 h after dosing. The animals were killed after 3 days and selected organs and tissues were removed. The amount of radioactivity remaining in the blood, selected tissues and the carcasses was estimated. Excretion of the radiolabel was rapid: in the first 24 h, in males, 18.8% of the dose was excreted in the urine and 68.3 % in the faeces. Equivalent figures for females were 10.3% and 70.7%. Of the radiolabel, 92.5% was eliminated within 72 h in the male rats and 93.9% in female rats. Tissue concentrations of radiolabel were generally lower in the females than in males. Only low concentrations of radiolabel were retained in the residual carcass (0.70% and 0.55% of the administered dose in males and females, respectively), the highest concentrations being found in the lungs, livers and kidneys (Lythgoe & Howard, 1995b).

In a third study of the absorption, distribution and excretion of paraquat, a single dose of ¹⁴C-labelled paraquat (50 mg of paraquat ion/kg bw) was administered by gavage to five male and five female Alpk: ApfSD rats. The specific activity of the dosing solution was 79.83 kBq/g. Urine was collected 6 h after dosing and urine and faeces were collected separately at 12, 24, 36, 48 and 72 h after dosing. The animals were killed after 3 days and selected organs and tissues were removed. The amount of radioactivity remaining in the blood, selected tissues and the carcasses was estimated. Excretion of the radiolabel was rapid: in the first 24 h, in males, 9.2% of the dose was excreted in the urine and 54.5 % in the faeces. Equivalent figures for females were 11.6% and 49.6%. Of the label, 92.7% was eliminated in 72 h in the male rats and 91.7% in female rats. The highest concentrations of radioactivity were retained in the lungs and residual carcass (Lythgoe & Howard, 1995 c).

Daniel & Cage (1966) investigated the absorption and excretion of paraquat (and diquat) in albino Wistar rats given ¹⁴C-labelled paraquat dichloride (0.94 mCi/mmol) as single oral doses at 4 or 6 mg/kg bw, or paraquat dimethosulfate as oral doses at 2.5-24 mg/kg bw, or subcutaneously at a dose of 21 or 23 mg/kg bw. Paraquat was poorly absorbed from the gut. After administration by either route, most of the radiolabel was found in the excreta within 2 days. After oral administration of paraquat, no radiolabel was detected in the bile (Daniel & Cage, 1966).

Dey et al. (1990) studied the pharmacokinetics of ¹⁴C-labelled paraquat (111 mCi/mmol) administered to male Sprague-Dawley rats as a single subcutaneous injection at a dose of 72 µmol/kg bw. This dose was considered to be one that would produce lung damage but avoid kidney damage. Blood was sampled through indwelling cannulae, and urine and faeces were collected at 2, 4, 6, 8, 12, and 24 h and then daily for 7 days. Non-cannulated rats treated in the same way were exsanguinated at intervals from 10 min to 7 days after dosing; tissue concentrations of ¹⁴C were measured in selected organs. The right lungs and kidneys were processed for histopathological examination. Histopathological examination showed changes characteristic of paraquat-induced lung pathology, without renal damage. Paraquat was rapidly absorbed, with peak blood concentrations of 58 µmol/l after 20min. The pharmacokinetics were best characterized as a two-compartment open model, the mean half-life (t½) being approximately 40 h. Highest tissue concentrations observed were in the kidney (358 nmol/g of tissue) and lung (64 nmol/g tissue), both at 40 min after administration of paraquat (Dey et al., 1990).

The distribution of paraquat in the brain was examined in male adult Wistar-derived Alderley Park rats after subcutaneous administration of paraquat (containing ¹⁴C-labelled paraquat with a specific activity 2 mCi/mmol) at a dose of 20 mg of ion/kg bw. The aim of this study was to determine whether paraquat crosses the blood-brain barrier. After administration, the concentration of radiolabel in the brain reached a maximum (0.05% of administered dose) within the first hour and then rapidly disappeared. Twenty-four hours after administration, however, a residual amount of paraquat still remained in the brain (1.6 nmol/g wet weight) and could not be removed by intracardiac perfusion. Most of the paraquat was associated with five structures, two of which (the pineal gland and linings of the cerebral ventricles) lie outside the blood-brain barrier. The remaining three brain areas (the anterior portion of the olfactory bulb, hypothalamus and area postrema) do not have a blood-brain barrier. Overall, the distribution of ¹⁴C-labelled paraquat in the brain 24 h after systemic administration was highly correlated to the blood volume. The authors concluded that paraquat remaining in the brain 24 h after systemic administration was associated with elements of the cerebral circulatory system, such as the endothelial cells that make up the capillary network, and that there was limited entry of paraquat into brain regions without a blood-brain barrier (Naylor et al., 1995).

The extent to which paraquat entered the brain was compared in groups of neonatal (aged 10 days), adult (aged 3 months) and elderly (aged 18 months) Wistar-derived Alpk: ApfSD rats. Both male and female neonatal rats were studied, while the adult and elderly rats were males. Groups of six to eight rats were given a single dose of [¹⁴C]paraquat (103 mCi/mmol) at 20 mg/kg, administered subcutaneously, and killed 30 min or 24 h after injection; blood was taken by cardiac puncture and brains were removed. Groups of four neonatal, adult or elderly rats were similarly injected and killed 24 or 48 h after dosing; the brains of these animals were subjected to histopathological examination. At all ages, plasma concentrations of paraquat were much higher at 30 min than at 24 h. At 30 min, the concentration of paraquat in the brain was highest in the elderly rats. While at 24 h the concentration in the brains of the adult and elderly rats had decreased, it remained high in the brains of the neonatal rats. Autoradiography showed similar distributions of paraquat in the various brain regions, paraquat being found in areas outside the blood-brain barrier or where the barrier is incomplete, e.g. the dorsal hypothalamus, area postrema and anterior olfactory bulb. There was no evidence for paraquat-induced cell damage in the neonatal brain, although there was increased paraquat entry into the brain in neonates than in older rats (Widdowson et al., 1996a).

In a study of the entry of paraquat into the brain, five male Wistar-derived Alpk: ApfSD rats were given paraquat (labelled with [¹⁴C]paraquat; specific activity, 20 µCi/ml) at a dose of 5 mg of ion/kg bw per day) daily for 14 days by oral administration, and another five rats received a single oral dose of paraquat (labelled with [¹⁴C]paraquat; specific activity 106 µCi/ml) at a dose of 5 mg ion/kg bw. The rats were killed 24 h after the last of the 14 doses or after the single dose. Concentrations of paraquat in the brain were 10 times higher in rats receiving multiple doses than in those receiving single doses. The same paper described a study of neuropathology, which included behavioural tests (see below) (Widdowson et al., 1996b).

In a study that used a brain microanalysis technique with detection by high-performance liquid chromatography-ultraviolet (HPLC-UV), paraquat, administered subcutaneously at a dose of 5, 10 or 20 mg/kg bw, was found to appear in the dialysate of the striatum in male Wistar rats. It was also found that paraquat did not allow 1,2,3,6-tetrahydropyridinium ion to penetrate the blood-brain barrier. Intraperitoneal injection of L-valine (200 mg/kg) 30 min before administration of paraquat at a dose of 20 mg/kg bw reduced the striatal extracellular concentrations of paraquat. The authors hypothesized that paraquat is taken up into the brain via the neutral amino acid transporter (Shimizu et al., 2001; see also McCormack & Di Monte, 2002).

In a study in anaesthetized male Wistar rats, the excretion of paraquat was found to be greater than the glomerular filtration rate, and to be concentration-dependent and saturable, implying that paraquat is secreted by a process involving active transport (Chan et al., 1997).

Groups of albino Wistar rats were given diets containing paraquat at a concentration of 50, 120 and 250 mg/kg (as paraquat ion) for 8 weeks. Groups comprised 30 animals at the two lower dietary concentrations and 40 animals at the highest concentration. After 2, 4 and 8 weeks, 10 rats per group were killed and selected organs were analysed for paraquat. At 50 mg/kg, paraquat was not detected in the kidneys, liver, brain or lung at any time, but was present in the gastrointestinal tract and, at low concentrations, in muscle. At 120 mg/kg, paraquat was detected in kidneys, lung and gastrointestinal tract. At 250 mg/kg, paraquat was detected in kidneys, lung and gastrointestinal tract (Litchfield et al., 1973).

Mice

The tissue distribution of paraquat was studied using whole body autoradiography in mice treated by intravenous injection. Mice received ¹⁴C-labelled paraquat at a dose of 20 mg of paraquat ion/kg. Two mice were killed at each time-point after the paraquat injection (10 min, 1, 5, 24 and 72 h). Paraquat was found to be concentrated in the liver and cartilage, and was not detected in the central nervous system. Paraquat was also present in the lungs, notably so after 24 h. At 72 h, radioabel was only present in the stomach and intestinal contents (Litchfield et al., 1973).

Hens

Three Warren hens were given gelatin capsules containing ¹⁴C-ring-labelled paraquat (purity, 99.7%; specific activity, 1.216 × 105 dpm/mg) at a daily dose of 4.52 mg of paraquat ion (0.247 mCi) for 10 days. One hen was used as the control. The hens were killed 4 h after the last dose. The highest concentration of radiolabel was found in the kidneys, while rather less was found in the gizzard and liver. Very little was found in fat. Paraquat was found at a concentration of 0.052 µg/g in eggs, mostly in the yolk (Hendley et al., 1976b).

Dogs

Greyhound dogs were given ¹⁴C-labelled paraquat at a dose of 30-50 µg/kg bw. The authors of this study considered that the kinetics could be described by a three-compartment open linear system (Bennett et al., 1976).

The elimination of paraquat was studied in the female greyhound dog. After intravenous injection of low doses (30-50 µg/kg) of ¹⁴C-labelled paraquat, radiolabel was found to be rapidly excreted in the urine, the clearance being greater than the glomerular filtration rate, suggesting a process of active secretion. Secretion could be inhibited by N’-nicotinamide. Large doses of paraquat (20 mg/kg bw) produced renal failure, as evidenced by a decrease in both renal creatinine and paraquat clearance. The elimination of paraquat could be described by a three-compartment open model (Hawksworth et al., 1981).

Goats

An English white nanny goat was dosed with ¹⁴C-ring-labelled paraquat (purity, 99.7%; specific activity, 2.28 × 104 dpm/µg) in the diet at a dose equivalent to 100 µg of paraquat ion/g of diet. This was done by adding 206.6 mg of radiolabelled paraquat (as ion) to the diet at the morning and afternoon feeding, daily for 7 days. Another nanny goat was used as the control. The goats were killed 4 h after the last feeding with radiolabelled paraquat. Radioactivity was measured in the urine, faeces, stomach, milk, and in selected tissues. At sacrifice, 2.4% and 50.5% of the administered material had been excreted in the urine and faeces respectively. The stomach contents included 33.2% of the administered dose. The highest concentration of radiolabel seen in the milk was 0.009 µg/g (on the morning of day 7). The highest tissue concentrations were found in the kidney and liver (Hendley et al., 1976a).

Pigs

A pig (Large White × Welsh boar) was given 100 mg of ¹⁴C-methyl-labelled paraquat (purity, 99.3%; specific activity, 4.88 × 104 dpm/µg of paraquat ion) on 7 consecutive days; this was calculated to be equivalent to about 50 µg of paraquat ion/g of diet. A second boar acted as the control. The daily dose was spotted onto the commercial pig diet. The pig was killed 2 h after the final dose. The highest concentrations of paraquat were present in the kidney and liver (Leahey et al., 1976).

In a second study in pigs, ¹⁴C-methyl-labelled paraquat dichloride (purity, 99.45%; specific activity, 4.72 dpm/µg) at a daily dose of approximately 100 mg of paraquat ion was administered twice daily for 7 days to a Large White × Welsh boar. The dose contained about 2 mCi of radiolabel and the content of paraquat was calculated to be equivalent to about 50 µg paraquat ion/g of diet. The daily dose was spotted onto commercial porcine diet pellets. A second boar acted as the control. The highest concentrations of radiolabel were found in the kidney, with somewhat less being found in the liver and lung (Spinks et al., 1976).

Monkeys

Purser & Rose (1979) studied the renal handling of paraquat administered orally at a dose of 85 mg of paraquat ion/kg bw (containing 500 µCi of ¹⁴C-labelled paraquat) to three male cynomolgus monkeys (Macaca fascicularis). In two monkeys, peak plasma concentrations were observed at 2 h in two monkeys and at 10 h in the third monkey. The renal clearance of paraquat was high during the first 10 h, but fell markedly as renal failure set in at 14 h. The clearance of paraquat was always well in excess of the clearance of creatinine, suggesting an active secretory process.

Studies in more than one species

The disposition of orally-administered ¹⁴C-labelled paraquat dichloride was studied in male Sprague-Dawley rats, male and female guinea-pigs, and monkeys (Macaca fascicularis). The doses used were: rats, 126 mg/kg bw (0.175 µCi/mg); guinea-pigs, 22 mg/kg bw (1.25 µCi/mg); and monkeys, 50 mg/kg bw (0.4 µCi/mg). In the case of the rats and guinea-pigs, the doses used were LD₅₀s at 7 days. A total of 61 rats, 21 guinea-pigs and three monkeys were used. For the rats and guinea-pigs, urine and faeces were collected and groups were sacrificed at various times up to 21 days after the administration of paraquat. Selected organs were collected at sacrifice. For the monkeys, blood samples were taken at 30 min, 1, 2, 4, 8, 16 and 32 h after administration of paraquat and daily thereafter. In the rats, deaths were seen mainly after 5 days. A large portion of the paraquat was not absorbed from the gastrointestinal tract. Peak serum concentrations of radiolabel were seen at 30-60 min, while concentrations of radiolabel were higher in liver, kidneys and lungs than in serum. Similar results were found in the guinea-pigs. In the monkeys, one of which died on day 8, serum concentrations of radiolabel decreased rapidly after the first time-point (Murray & Gibson, 1974).

There is evidence that paraquat is taken up into the lungs by a process of active uptake, the normal substrate being endogenous diamines, e.g. putrescine and polyamines such as spermine and spermidine (see review by Smith, 1985). Diquat is not a substrate for this system and this fact accounts for the different organ-specific toxicity of these two bipyridilium herbicides (this is discussed further below).

1.2 Biotransformation

Rats

In the Daniel & Cage (1966) study in albino Wistar rats treated with ¹⁴C-labelled paraquat dichloride, discussed above, some evidence of metabolism was found. Of the dose of paraquat administered orally, 30% of the radiolabel was present in the gut as metabolic products. Furthermore, a small amount of metabolite was present in the urine after oral but not subcutaneous administration, suggesting that metabolites were absorbed from the gut. Studies in vitro, using faecal homogenates, suggested that microbiological metabolism was responsible for this. In the study of Murray & Gibson (1974) in male Sprague-Dawley rats, male and female (mixed) guinea-pigs and cynomolgus monkeys (Macaca fascicularis), metabolites were not observed.

Urine and faeces samples from the studies in rats, described above (Lythgoe & Howard, 1995a, b, c), were pooled separately for the females and males of each study for the whole 72 h of that study. After extraction, samples were analysed by thin-layer chromatography. In all cases, paraquat accounted for the vast majority of the radiolabel in the urine (95.0% of urinary label in the males receiving the lower dose and 93.6% females receiving the lower dose). Three minor metabolites were found in urine; these were not further identified. Faecal material showed that the vast majority of the radioactivity in all cases was unchanged paraquat. It was therefore concluded that paraquat was largely unmetabolized (Macpherson, 1995).

Hens

In the study in hens, residues in tissues were generally unchanged paraquat. A small amount of a metabolite, l-methyl-(4'-pyridyl), was found in the livers and kidneys (Hendley et al., 1976b).

Goats

In the study in goats, residues in tissues were generally unchanged paraquat. In the liver, small amounts of 4-(1,2-dihydro-1-methyl-2-oxo-4-pyridyl)-1-methyl pyridinium ion and 1-methyl-4-(4'-pyridyl) pyridinium ion were found. The latter compound was also found in peritoneal fat (Hendley et al., 1976a).

Pigs

In the study by Leahey et al. (1976), all the radiolabel in the tissues, except the liver, was found to be in the form of paraquat. In the liver, 7% of the radiolabel was accounted for by l-methyl-4-(4'-pyridyl) pyridinium ion. In the study by Spinks et al. (1976), 4% of the radiolabel was accounted for by l-methyl-4-(4'-pyridyl) pyridinium and 70% by unchanged paraquat.

2.1 Acute toxicity

The results of studies on the acute toxicity of paraquat administered by a variety of routes are summarized in Table 1. Clinical signs seen after administration of paraquat by the oral, subcutaneous or intraperitoneal routes included decreased activity, dehydration and breathing irregularity. In animals that died after administration of paraquat by these routes, mottled areas of lung were seen. Scabbing of skin was seen after administration by the dermal route, but no systemic signs of poisoning were present. After rats had inhaled paraquat, clinical signs and appearances post mortem were similar to those seen after oral, subcutaneous or intraperitoneal administration.

Table 1. Acute toxicity of paraquat

NS, not stated; M, male; F, female
^a Paraquat dichloride; purity, >98%
^b Dose quoted as paraquat ion
^c Technical paraquat dichloride (33% w/w paraquat ion)
^d as dimethylsulfate
^e LC₅₀ (at 4 h) (mg of paraquat ion/m^-3)
^f Material used was paraquat dichloride, 21.5% w/v, but results were expressed as paraquat ion; aerosol mass median aerodynamic diameter (MMAD), <0.3 µm; rats exposed by nose only

(a) Dermal irritation

The dermal irritation potential of paraquat dichloride technical concentrate (paraquat ion, 33% w/w) was assessed in young adult female New Zealand white albino rabbits. Undiluted test material was applied to the depilated left flank of the rabbits, which was then covered by gauze and impermeable rubber. These were left in place for 4 h. After removal of the dressing and cleansing of the application site, the Draize scale was used to assess erythema and oedema, 30-60 min and 1, 2, 3, 4, 7, 14, 17, 20, 21 and 23 days after exposure. Slight erythema was observed, which regressed by 3 days and 4 days in two animals, but still remained after 23 days in the third animal. Very slight transient oedema was seen in one animal (at the 30/60 min observation time), very slight oedema was seen in the second, this still being present at 4 days but not at 7 days, while there was no oedema in the third rabbit (Duerden, 1994c).

(b) Ocular irritation

The potential for paraquat dichloride technical concentrate (paraquat ion, 33% w/w) to produce irritation of the eye was assessed in young adult female New Zealand white albino rabbits. Test material (0.1 ml) was applied to the left eye of each rabbit. Rabbits were dosed sequentially; and mild systemic toxicity was noted in the third rabbit to be dosed. Accordingly, this rabbit was killed. The fourth rabbit was collared to prevent oral ingestion of the test material. The eyes of rabbits 1, 2 and 4 were then examined and the degree of irritation was assessed using the Draize scale from 1 h to up to 28 days after instillation. Initial pain was graded as slight or was absent. Slight or mild corneal opacity was seen in all three animals, this effect resolving within 17 days. Redness and chemosis of the conjunctiva was seen in all animals and resolved by 28 days and 14 days after exposure. No effect was seen on the iris, while erythema of the eyelids and mucoid discharge was observed. Paraquat was considered to be a moderate ocular irritant (Bugg & Duerden, 1994).

In a study of ocular toxicity, paraquat was administered at a concentration of 6.25, 12.5, 25, 50 and 100% of a solution containing 242 mg of paraquat ion/ml. A total of 15 male New Zealand white rabbits were used, nine rabbits receiving different doses in each eye and six rabbits receiving the same dose in both eyes. Control eyes received normal saline. In all cases, 0.2 ml of solution was pipetted into the lower conjunctival sac, and the eyes were examined at 12 h and then daily for 20 days, ocular lesions being scored on the Draize scale. At 6.25 and 12.5%, severe conjunctival reactions were seen, with occasional slight corneal damage at 12.5%. At higher concentrations (25 and 50%), the iris was congested and swollen and there was corneal opacification; a pannus reaction was also seen. Animals to which the 100% solution was administered died within 6 days. The time of maximal effect was around 9 days and those who received the 25% and weaker solutions showed recovery thereafter (Sinow & Wei, 1973).

(c) Dermal sensitization

A study of the sensitization potential of paraquat dichloride technical concentrate (paraquat ion, 33% w/w) was based on the maximization test of Magnusson & Kligman (1969). Female albino (Hsd/Poc: DH) guinea-pigs were used. The positive control was 2-mercaptobenzothiazole. A preliminary study was carried out to determine the concentrations of test material that gave an acceptable degree of irritancy and no signs of systemic toxicity. In the main study, 30 guinea-pigs were used (20 as test animals and 10 as controls). For induction, each animal received Freund complete adjuvant diluted 1:1 with deionized water, 0.03% w/v test material, and 0.03% w/v test material with Freund complete adjuvant 1:1 with deionized water, which were injected intradermally at three different sites in the previously depilated scapula region. One week later, the scapula region was again clipped and the test material (10% w/v) was applied topically over the injection sites. Animals serving as negative controls were treated in the same way except that the three inducing injections were Freund complete adjuvant 1:1 with deionized water, deionized water, and again Freund complete adjuvant 1:1 with deionized water. Animals serving as positive controls (20) were treated in the same way as the test animals except that the test substance administered was 2-mercaptobenzothiazole, and there were 10 negative controls for this group. For these guinea-pigs, the topical applications consisted of deionized water. Two weeks after the topical applications, both flanks of all animals were clipped free of hair and a preparation of 30% test material on an occlusive dressing was applied to one flank and a preparation of 10% test material to the other flank. These were left in place for 24 h. Erythematous reactions were recorded at 24 h and 48 h later. One animal in the test group died, but no erythema was found at either time in this group, nor in the negative control group. In contrast, erythema was seen in 19 of the positive controls, and it was concluded that paraquat had no sensitization potential (Duerden, 1994d).

2.2 Short-term studies of toxicity

(a) Oral administration

Mouse

In a 13-week dietary study, groups of 20 male and 20 female ICR-CRJ SPF mice were given paraquat dichloride (purity, 93.3%) at a dietary concentration of 0, 10, 30, 100 and 300 mg/kg, equal to 1.18, 3.65, 11.5 and 35.8 mg of paraquat dichloride/kg bw per day in males and 1.38, 3.91, 13.8 and 41.9 mg of paraquat dichloride/kg bw per day in females. These doses are equal to 0, 0.85, 2.64, 8.33 and 25.9 mg of paraquat ion/kg bw per day in males and 0, 1.00, 2.83, 9.99 and 30.3 mg of paraquat ion/kg bw per day in females. Mice were observed daily for mortality and daily clinical observations were undertaken. Animals found dead or that were killed in extremis were subjected to immediate autopsy. The mice were weighed weekly and food and water consumption were measured twice per week. On day 91, blood was collected from at least 10 mice from each group for haematological examination and for clinical chemistry. The mice were then examined post mortem. Autopsy was carried out on the remainder of the mice the next day, at which time urine was collected for urine analysis. At necropsy, selected organs were weighed and these and other organs were fixed and sections made for histopathological examination. Mortality was observed at 300 mg/kg, two females dying from pulmonary damage, one in week 2 and one in week 11. The decedents' lungs showed pulmonary oedema, small round cell infiltration with phagocytosis, and, in one animal, eosinophilic swelling of the epithelial cells of the alveoli. At 300 mg/kg in both sexes, there was reduced body-weight gain, almost from the inception of the study, however, these values were only significantly different from those of controls at a few time intervals. No intergroup difference in food intake was observed, but a slight reduction in food conversion efficiency was seen at 300 mg/kg in both sexes. No intergroup differences were seen in water intake. No test material-related intergroup differences were seen in haematological parameters (although a reduction in mean corpuscular volume at 300 mg/kg may have been test material-related in males) or in clinical chemistry findings. Terminal body weights were reduced in males at the highest dietary concentration, as were the absolute weights of the heart, liver and muscle. An increase in relative lung weight and a decrease in relative liver weight were also seen. In females at the highest dietary concentration, an increase in absolute pituitary, lung, kidney and spleen weight was observed, accompanied by a decrease in ovarian weight. Relative weights of organs also showed some intergroup differences, those of the pituitary, thyroids, lung, kidneys and spleen being increased while that of the ovaries was decreased. Eosinophilic hypertrophy of alveolar epithelial cells was observed in both sexes (17 out of 20 males, and 12 out of 18 surviving females). Pulmonary oedema and alveolar cell proliferation was also seen in a few males and in one female. Accordingly, the no-observed-adverse-effect level (NOAEL) for the study was 100 mg/kg (equal to 11.5 mg of paraquat dichloride/kg bw per day for males and 13.8 mg of paraquat dichloride/kg bw per day for females), on the basis of decreased body-weight gain and histopathological changes in the lungs at 300 mg/kg. These NOAELs are equal to 8.33 mg of paraquat ion/kg bw per day in males and 9.99 mg of paraquat ion/kg bw per day in females (Maita et al., 1980a).

Rat

In a 13-week study, groups of 20 male and 20 female Fischer CDF (F344) CRJ rats were given diets containing paraquat dichloride (purity, 93.8%) at a concentration of 0, 10, 30, 100 or 300 mg/kg (0, 7, 22, 72 and 217 mg/kg of paraquat ion, equal to 0, 0.49, 1.44, 4.74, 14.2 mg of paraquat ion/kg bw per day for males and 0, 0.52, 1.53, 5.14 and 15.27 mg of paraquat ion/kg bw per day for females). Another group received diet containing no test material and acted as controls. The rats were examined daily for adverse clinical signs, body weight was measured weekly and food and water intake were measured twice per week. Ninety-one days after the start of the study, at least 10 animals of each sex per group were chosen for blood sampling. The samples were used for haematology and clinical chemistry and, after sampling, the animals were examined post mortem. On day 92, urine analysis was carried out on the remaining rats, which were examined post mortem. At necropsy, selected organs were weighed and these and other organs were fixed in 10% formalin; they were then processed for histopathological examination. No rats died during the study and no abnormal clinical signs were seen. At the highest dietary concentration there was markedly reduced body-weight gain and decreased food and water intake in both sexes. Neither reduced body-weight gain nor reduced food intake was seen at lower dietary concentrations. No test material-related abnormalities were found on haematological examination, clinical chemistry or urine analysis. In the males, terminal body weights and absolute weights of brains, pituitaries, thyroids, livers, kidneys, spleens and muscles were decreased at the highest dietary concentration. Also in males, relative weights of brain, pituitary, lung, kidneys, adrenals, testes and muscle were increased. In females, terminal body weights were depressed at the highest dietary concentration, together with the absolute weight of the heart, lung and liver. Relative brain, lung, kidney, ovary and muscle weights were increased at the highest dietary concentration. These changes probably reflected the reduced food intake at the highest dietary concentration. On histopathological examination, alveolar epithelial hypertrophy was observed in males (6 out of 20) while in females, there was an increased prevalence of brown pigmentation of the spleen, both at the highest dietary concentration. The NOAEL was therefore 100 mg/kg in both sexes, equal to 4.74 mg of paraquat ion/kg bw per day for males and 5.14 mg of paraquat ion/kg bw for females on the basis of reduced body-weight gain and reduced food and water intake at the highest dietary concentration, together with pathological changes in the lungs and spleen (Maita et al., 1980b).

Dog

In a 6-week study, groups of three male and three female beagle dogs received technical-grade paraquat (purity, 32.2%) at a dietary concentration of 35 or 90 mg/kg as paraquat ion (equivalent to 0.875 and 2.25 mg of paraquat ion/kg bw per day) for 6 weeks. An additional group of three males and three females received capsules containing paraquat at a dose of 0.75 mg of paraquat ion/kg bw per day, also for 6 weeks. The controls from Sheppard (1981b) were used (see below) and the results were also compared with the group receiving paraquat at 20 mg/kg in that study, as this is comparable to the dose of 0.75 mg/kg bw per day in capsules. Animals were observed periodically during the working day, and by a veterinarian before the start of the study and preterminally. Ophthalmoscopy and auscultation of the chest were undertaken before the start of the study and before termination. Body weights were recorded weekly and food consumption was measured daily. Blood was taken for clinical pathology before the start of treatment, and after 3 and 5 weeks of treatment. Urine analysis was carried out. Lungs and kidneys were weighed at necropsy, and these organs and portions of other selected organs were processed for histopathological examination. There were no adverse clinical effects, nor were there any paraquat-related effects on ophthalmoscopy. On auscultation, increased respiratory sounds were heard in animals from several groups: of these, the finding in two males and three females at 90 mg/kg may have been test material-related. Body weights decreased in males at a dietary concentration of 90 mg/kg throughout the study and in the females at 90 mg/kg towards the end of the study. Body-weight gain was reduced in those females given paraquat at 0.75 mg/kg bw per day. Food intake was reduced at 90 mg/kg in females towards the end of the study. In the males fed paraquat at a dietary concentration of 90 mg/kg, there was a reduction in erythrocyte volume fraction, haemoglobin and erythrocyte count. No test material-related findings were seen in clinical chemistry investigations or urine analysis. One female fed the diet containing paraquat at 90 mg/kg had a markedly increased lung weight. Changes were seen at 0.75 mg/kg bw per day and at 90 mg/kg in the lungs, both macroscopically and microscopically. The macroscopic changes comprised grey, red or purple depressed areas. In all animals receiving capsules containing paraquat at 0.75 mg/kg bw per day, and in five of the six animals receiving diet containing paraquat at a concentration of 90 mg/kg, there was histopathological evidence of alveolitis, such as intra-alveolar accumulations of mononuclear cells, interstitial hypercellularity and fibrosis and alveolar hyperplasia. Occasional giant cells and pigmented macrophages were seen. It was concluded that for paraquat administered in the diet the NOAEL was 35 mg/kg (equivalent to 0.875 mg of paraquat ion/kg bw per day) and that paraquat was more toxic when administered by capsule than when mixed with the diet (Sheppard, 1981a).

In a 13-week study, groups of three male and three female beagle dogs received paraquat (paraquat ion, 32.2% w/w) at a dietary concentration of 0, 7, 20, 60 or 120 mg/kg as paraquat ion. These dietary concentrations are equal to doses of 0, 0.20, 0.55, 1.75 and 3.52 mg of paraquat ion/kg bw per day in males and 0, 0.24, 0.71, 1.92 and 4.26 mg of paraquat ion/kg bw per day in females. Animals were observed more than once daily, and by a veterinarian before the start of treatment and after 6 and 12 weeks of treatment. Ophthalmoscopy was carried out before the start of the study and after 6 weeks of treatment. Auscultation of the chest was carried out before the start of the study, 6 weeks after the start and immediately before the end of the study. The animals were weighed weekly and food consumption was measured daily. Blood samples were taken by jugular venepuncture before the start of treatment and after 6 and 12 weeks of treatment. These samples were used for haematological investigations and for clinical chemistry. At autopsy, selected organs were weighed and these and other selected organs were fixed and processed for histopathological examination. At the highest dietary concentration two males and two females were killed in extremis. These animals exhibited marked dyspnoea as well as increased respiratory sounds (harsh râles) and loss of body weight before they were killed. One female at 200 mg/kg showed pyrexia and malaise at 3 weeks; this was treated successfully with procaine penicillin and dihydrostreptomycin. The same animal showed loss of appetite from week 8; it was again treated with antibiotics. Survivors at the highest dose showed body-weight loss. Slight but significant reductions in weight gain were seen in females at 7, 20 and 60 mg/kg, compared with the controls; there was no clear dose-response relationship. These effects were not considered to be related to treatment. Food consumption was reduced in one female at the highest dietary concentration. Retinal engorgement was seen in one animal each at 7 mg/kg and 20 mg/kg, and two at 120 mg/kg. No intergroup treatment-related haematological or clinical chemistry findings were present, except in one of the decedents where haemoconcentration was seen. No test material-related effects on urinary parameters were seen. Absolute and relative lung weights were increased in all animals at 120 mg/kg and in two animals at 60 mg/kg; while not statistically significant (lungs from only two animals of each sex were weighed at the highest dietary concentration), these findings were considered to be biologically significant. Histopathological changes were seen in the lungs at 60 and 120 mg/kg. These changes consisted of proliferative alveolitis, with interstitial cellular infiltration (eosinophils and polymorphs) and exudate. Some renal (distal tubular) changes were seen in the same groups. The NOAEL was considered to be 20 mg/kg, equal to 0.55 mg/kg of paraquat ion per kg bw per day in males, and 0.71 mg/kg of paraquat ion per kg bw per day in females, on the basis of increases in lung weight and histopathological changes at the next higher dietary concentration (Sheppard, 1981b).

In a 1-year feeding study, groups of six male and six female beagle dogs were given diets containing technical-grade paraquat dichloride (paraquat ion, 32.2%) at a concentration of 0, 15, 30 or 50 mg/kg as paraquat ion for 1 year. Although no overall means were given in the study report, they were quoted in the submission document. Intakes were 0, 0.45, 0.93 and 1.51 mg of paraquat ion/kg bw per day for males and 0, 0.48, 1.00 and 1.58 mg of paraquat ion/kg bw per day for females (see Clapp, 2002). The dogs were observed twice daily, and examined by a veterinarian before the start of the study and after 13, 26 and 39 weeks, and also between weeks 48 and 51 of treatment; the examination by the veterinarian included auscultation and ophthalmoscopy. Body weights were measured weekly and food consumption daily. Haematology and clinical chemistry measurements were carried out during the study on jugular venous blood samples taken before the start of the study and at weeks 4, 8, 12, 16, 20, 39 and 52. Urine for urine analysis was collected over an 18 h period before the start of the study, and at weeks 8, 16, 24, 39 and 50. Urine samples were collected at week 39 for analysis for paraquat. At termination, necropsy was undertaken and selected organs were weighed, and these and other organs were processed for histopathological examination. Samples of kidney, liver and lung, taken at necropsy, were analysed for paraquat. Respiratory dysfunction was observed at 50 mg/kg (hyperpnea). Increased vesicular sounds were heard in the lungs at auscultation. Erythema of the dorsum of the tongue was seen at 30 and 50 mg/kg in males, and at 50 mg/kg in females. No test material-related effects were seen on ophthalmoscopy. No test material-related effect on body-weight gain was seen. Reduced food consumption was seen at 50 mg/kg. No haematological changes were seen that were attributable to paraquat. Alkaline phosphatase activity was elevated in females at 30 and 50 mg/kg, and plasma concentrations of triglycerides were raised in both sexes at 50 mg/kg. Urinary specific gravity was elevated at 50 mg/kg in males. Lung weights (both absolute and relative) were significantly increased at 50 mg/kg. Spleen weights were elevated at 50 mg/kg in both sexes, but the biological significance was unclear, and the mean in males was influenced by one outlier. At 30 and 50 mg/kg, macroscopically there was yellow discoloration in the lungs. Microscopically, there was peribronchial mononuclear infiltration, peribronchial and interalveolar fibrosis and changes in the alveolar epithelium (alveolar cell hyperplasia and hypertrophy). These changes were accompanied by the presence of haemosiderin-laden macrophages. These changes were more severe at 50 mg/kg than at 30 mg/kg. Erythrophagocytosis in the bronchial lymph nodes was present at 30 mg/kg and 50 mg/kg. A dose-related increase in urinary paraquat was found at week 29. Paraquat was not found in the livers at any dietary concentration, but was found in the kidneys at 30 and 50 mg/kg. Paraquat was detected in the lungs. The NOAEL for the study was 15 mg/kg on the basis of erythema of the tongue at 30 mg/kg in males, elevated alkaline phosphatase in both sexes, and histopathological changes in the lung at > 30 mg/kg. This NOAEL is equal to 0.45 mg of paraquat ion/kg bw per day (Kalinowski et al., 1983).

Cows

Groups of two Friesian cows were fed diets containing paraquat (as residues in dried grass) at a concentration of 25, 80 or 170 mg/kg as paraquat ion for 3 months. These dietary concentrations were equivalent to 0.375, 1.2 and 2.55 mg of paraquat ion/kg bw per day. During the trial, milk was collected from the cows. After they had been slaughtered, autopsy was carried out and organs, inter alia lung, liver and kidney, were examined histopathologically. Concentrations of paraquat were determined in the liver, kidney, renal fat and the pectoralis and adductor muscles. No adverse clinical effects were noted during the study, although the milk yield decreased (this was attributed to poor nutrition). No histopathological change attributable to paraquat was found. The concentration of paraquat in the milk was very low (in one sample, 0.001 mg/kg; in the remainder, <0.0006 mg/kg). The highest tissue residues were in the kidney (0.20-0.31) and liver (<0.01-0.09). Concentration in cardiac and skeletal muscle and renal fat were much lower. The NOAEL was the highest dietary concentration, 170 mg/kg, equivalent to 2.55 mg of paraquat ion/kg bw per day (Edwards et al., 1974).

(b) Dermal application

Rabbits

In a 21-day study of dermal toxicity, groups of six male and six female New Zealand white albino rabbits were given technical-grade paraquat (purity, 33.5%), at a dose of 0, 1.5, 3.4, 7.8 or 17.9 mg/kg bw per day (equal to 0, 0.5, 1.15, 2.6 and 6.0 mg of paraquat ion/kg bw per day), applied in distilled water under an occlusive dressing to the clipped dorsal thorax. Distilled water without paraquat was applied to the control animals. The period of exposure was 6 h per day. Animals were observed twice daily. They were more thoroughly examined and dermal irritation was assessed on days 1, 2, 4, 8, 11, 15, 18 and 21. Animals were weighed twice weekly and food consumption was measured weekly. Blood samples for haematology and clinical chemistry were collected before the start of the study and at termination. After 21 days, the animals were weighed and killed, and designated organs were weighed. These and further selected tissues were fixed and processed for microscopic pathological examination. No mortality was observed and all animals appeared to be clinically normal throughout the study. Body weights and food consumption were similar in all groups. No differences between the groups were seen in haematological measurements or clinical chemistry. Neither organ weight data nor histopathological examination showed evidence of test material-related systemic toxicity. Evidence of skin irritation was seen at the two highest doses. Microscopic evidence of skin irritation was seen in most animals at the highest dose and in some animals at a dose of 2.6 mg of paraquat ion/kg bw per day. Findings included erythema, erosion, ulceration, exudate, acanthosis and chronic inflammatory change. Accordingly, the NOAEL was 1.15 mg of paraquat ion/kg bw per day on the basis of skin changes at higher doses (Cox, 1986).

(c) Exposure by inhalation

Rats

In a 3-week inhalation study, an aerosol of technical-grade paraquat (paraquat ion, about 40%) was administered to groups of 36 male and 32 female albino Sprague-Dawley CD rats The rats were exposed for 6 h per day, 5 days per week, for 3 weeks (i.e. 15 exposures). There were two control groups, one of which received no exposure to aerosol and the other received a saline aerosol. There were two test groups, one of which received aerosolized paraquat at a concentration of 0.01 µg of paraquat ion/l and the other 0.1 µg of paraquat ion/l. Particles had aerodynamic diameters of <0.7 µm. The rats were examined twice daily and, more thoroughly, once a week. Animals were weighed daily for the first week and then twice per week. Food consumption was measured weekly. Water consumption was measured daily, 5 days per week. Interim kills were carried out as follows: 3 days after the first exposure, four males and four females in each group were killed for histopathological examination of the nasal passages, pharynx, larynx and lungs (i.e. the rats were given a single exposure, left for 2 days, and then sacrificed). Additionally, 1 day after the third exposure, four males in each group were killed for examination of the nasal turbinates only. Eight animals of each sex per group were killed after the last exposure and the remainder (16 animals of each sex per group) were killed after a 3-week recovery period. Macroscopic examination was carried out post mortem but no microscopic pathology was performed. No treatment-related clinical signs were seen. No treatment had any effect on body-weight gain or food or water consumption. Aerosol containing paraquat ion at a concentration of 0.01 µg/l did not produce histopathological changes in the larynx, while exposure to aerosol containing paraquat ion at a concentration of 0.1 µg/l did produce such changes. In the animals examined 3 days after exposure at 0.1 µg of paraquat ion/l, there was squamous metaplasia at the base of the epiglottis. One day after the third exposure, there was ulceration, necrosis, acute inflammatory change and squamous metaplasia and/or hyperplasia especially at the base of the epiglottis and arytenoid processes. In those animals sacrificed in the interim for examination of the turbinates, no abnormalities were seen. Accordingly, the NOAEL for the study was 0.01 µg of paraquat ion/l on the basis of histopathological changes in the upper respiratory tract at the higher dose (Grimshaw et al., 1979).

2.3 Long-term studies of toxicity and carcinogenicity

Rats

In a 104-week study, groups of 80 male and 80 female Fischer (F344/DuCrj) rats were given diets containing paraquat dichloride (purity, >98%) at a concentration of 0, 10, 30, 100 or 300 mg/kg. Intakes of paraquat dichloride were 0, 0.35, 1.06, 3.52 and 10.6 mg/kg bw per day for males, and 0, 0.43, 1.34, 4.32 and 11.7 mg/kg bw per day for females. These intakes of paraquat dichloride represented intakes of 0, 0.26, 0.77, 2.55 and 7.67 mg of paraquat ion/kg bw per day in males, and 0, 0.31, 0.97, 3.13 and 8.47 mg of paraquat ion/kg bw per day in females. Eight rats of each sex were killed at 26, 52 and 78 weeks, while the survivors were sacrificed at 104 weeks. During the study, animals were observed daily and clinical findings, including mortality, were recorded. Animals that died during the study were subjected to necropsy followed by histopathological examination, as were those that were sacrificed in extremis. Body weight was measured weekly until week 26 of the study, and fortnightly thereafter. Food and water consumption was measured twice per week. At termination, haematological and clinical chemistry studies were carried out on 10 males and 10 females per group. At necropsy, selected organs were weighed and portions of these and of other organs were fixed and processed for histopathological examination. No clinical effect attributable to the test material was seen, but there was some indication of increased mortality between weeks 66 and 74 of the study in females at the highest dose. There was a reduction in body-weight gain and food and water consumption in both sexes at 300 mg/kg (the highest dietary concentration). The effect on body-weight gain was greater in the males. Some effects on haematology were observed. At 26 weeks, there was a decrease in white blood cell count at 300 mg/kg in males, but no differences between groups were seen in females. At 52 weeks, there were minor changes in mean corpuscular haemoglobin and haemoglobin concentration, and a decrease in white blood cell count at 300 mg/kg in males, but no differences between groups were seen in females. At 78 weeks, there was a decrease in white blood cell count at 300 mg/kg in males, but no test material-related differences between groups were seen in females. At 104 weeks, in males, there were minor changes in mean corpuscular volume and mean corpuscular haemoglobin. In females, at 104 weeks, there were minor changes in mean corpuscular haemoglobin concentration (a reduction) at 300 mg/kg. At 26 weeks, in males, a reduction in aspartate aminotransferase activity and globulin was observed at 300 mg/kg, as well as a rise in blood concentrations of glucose at 100 and 300 mg/kg. At 26 weeks, in females, an increase was seen in gamma-glutamyl transpeptidase at 300 mg/kg, and a decrease at 10 mg/kg. At 26 weeks, total protein, albumin and globulin concentrations were all decreased in females at 300 mg/kg. At 52 weeks, in males, a reduction in aspartate aminotransferase, alanine aminotransferase and gamma-glutamyl transpeptidase activity, and in cholesterol and calcium concentrations was seen at 300 mg/kg. At 52 weeks, females showed no test material-related changes in clinical chemistry. At 78 weeks, males showed reductions in alkaline phosphatase, alanine aminotransferase and gamma-glutamyl transpeptidase activity were seen accompanied by an increase in albumin and a decrease in globulin at 300 mg/kg, whereas females showed no test material-related changes in clinical chemistry. At termination at 104 weeks, a decrease in globulin was seen in males at 300 mg/kg, while no noteworthy changes in clinical chemistry parameters were seen in females. In males at the highest dietary concentration, body weight at necropsy was decreased at 26, 52 and 78 weeks and at termination. Although some changes in organ weights were observed, many of these did not appear to be test material-related. At 26 weeks in males at 300 mg/kg, however, relative but not absolute lung weight was increased, as it was at 52 weeks. At 300 mg/kg, at 78 weeks, absolute lung weight was decreased and relative lung weight in males did not differ from those of controls, while at termination, neither value was different from that of controls. In females, at 26 weeks, a reduction in body weight was not seen at any concentration. At 26 weeks, an increase in relative lung weight was seen at the two higher dietary concentrations, and this was accompanied by an increase in absolute lung weight at 100 mg/kg only. At 52 weeks and 78 weeks, there were no differences between groups in body weight or absolute or relative lung weight in females. At termination, in females, decreased body weight and an increased relative but not absolute lung weight was observed at 300 mg/kg. A reduction in absolute and relative ovarian weight was observed at 26 weeks at the highest dietary concentration. On histopathological examination, there were changes in the lungs at 300 mg/kg in both sexes and at 100 mg/kg in males. The changes consisted of proliferation of interalveolar septum cells and hyperplasia of the alveolar epithelium. The frequency of pulmonary adenoma was increased in females at 300 mg/kg (see Table 2). Histopathological evidence of cataract was found in both sexes at 300 mg/kg (see Table 3). The NOAEL for the study was 30 mg/kg (1.06 mg of paraquat dichloride/kg bw per day and 1.34 mg of paraquat dichloride/kg bw per day in males and females respectively) on the basis of clinical chemistry changes in males, increased lung weight in females and histopathological changes in the lungs of males at >100 mg/kg. These NOAELs are equal to 0.77 and 0.97 mg of paraquat ion/kg bw per day in males and females respectively (Yoshida et al., 1982).

Table 2. Incidence of lung tumours in rats fed diets containing paraquat (survivors + decedents)

From Yoshida et al., (1982)

Table 3. Incidence of cataract in rats fed diets containing paraquat (decedents + survivors)

From Yoshida et al. (1982)

In a study of chronic toxicity, groups of 62 male and 62 female JCL: Wistar rats were fed diets containing paraquat (purity, 98%) at a concentration of 0, 6, 30, 100 or 300 mg/kg for up to 104 weeks. These dietary concentrations provided intakes of paraquat di chloride equal to 0, 0.25, 1.26, 4.15 and 12.25 mg/kg bw per day in males, and 0, 0.3, 1.5, 5.12, 15.29 mg/kg bw per day in females. These intakes are equal to 0, 0.18, 0.91, 3.00 and 8.87 mg of paraquat ion/kg bw per day in males, and 0, 0.22, 1.09, 3.71 amd 11.1 mg of paraquat ion/kg bw per day in females. Six rats of each sex per group were killed at 26 weeks and 52 weeks; the survivors were killed at 104 weeks. The rats were examined twice per day, deaths were recorded and clinical findings noted. Ophthalmoscopy was carried out before treatment, and before sacrifice in those killed at 26 and 52 weeks, and at termination. Body weight and food consumption were measured weekly until week 26 and thence fortnightly. Haematological and clinical chemistry end-points were measured in blood samples taken from animals killed at 26 weeks, at 52 weeks, and from the survivors at termination. Included among the clinical chemical parameters measured were activities of plasma, erythrocyte and brain cholinesterases. Urine analysis was performed on the animals killed at 26 and 52 weeks and on the survivors at termination. Animals killed at 26 and 52 weeks and survivors to termination were subjected to necropsy, as were decedents. Selected organs were weighed and these and other organs were fixed and processed for histopathological examination. No clinical effects were observed. At the highest dietary concentration in females, there was a decrease in weight gain in the middle of the study (weeks 43, 42-48 and 54), otherwise body-weight gain was unaffected. No substantial intergroup differences in food consumption or in water intake were noted. At week 26, at the highest dietary concentration, there was a decrease in erythrocyte count, erythrocyte volume fraction and haemoglobin and a reticulocytosis in males and in the erythrocyte count and haemoglobin in females. At 300 mg/kg, at week 52, decreased erythrocyte count and increased numbers of polymorphs were seen in males, and lowered erythrocyte count, haemoglobin concentration and leukocytes were seen in females. At week 104, both sexes showed decreases in erythrocyte count, erythrocyte volume fraction and haemoglobin, and an increase in reticulocytes was observed in males. At 26 weeks, a decrease in total protein was seen in both sexes at 300 mg/kg, with a decrease in alkaline phosphatase activity in females at this dietary concentration. At week 52, decreased total protein was found in both sexes, as well as reduced blood concentrations of glucose in males and reduced aspartate aminotransferase and alanine aminotransferase activities in females. There were no differences between the groups in urine analysis at any time-point. In the rats sacrificed at week 26, there were increases in absolute kidney weights (right kidney only) in males and in both absolute kidney weights in females and absolute ovarian weights in females. At week 52, at the highest dietary concentration, in males there was an increase in both absolute thyroid and kidney weights, while females showed an increase in absolute ovarian weights and a decrease in the relative weights of the brain, heart and liver. At termination, at 300 mg/kg, males showed reductions in the absolute and relative heart weights, while females showed lowered absolute and relative liver weights and decreased absolute heart weight. At necropsy and histopathological examination, no findings could be attributed to the test material. The NOAEL was therefore 100 mg/kg in both sexes (equal to 4.15 and 5.12 mg of paraquat dichloride/kg bw per day in males and females respectively), on the basis of haematological observations and lowered plasma concentration of total protein at the highest dietary concentration. These NOAELs are equal 3.00 and 3.71 mg of paraquat ion/kg bw per day in males and females, respectively (Toyoshima et al., 1982).

Groups of 70 male and 70 female Fischer 344 rats were given diets containing paraquat (technical grade, 32.69%) at a dietary concentration of 25, 75 or 150 mg/kg as paraquat ion (equivalent to a dose of 1.25, 3.75 or 7.5 mg of paraquat ion/kg bw per day) for a period of at least 113 weeks (males) and 122 weeks (females). Two additional groups of rats served as controls. There were also additional satellite groups of five males and five females from one control group and all three test groups which were sacrificed at 1 year for estimation of paraquat concentrations in certain tissues. Ten male and ten female rats from each group were sacrificed for histopathological examination at 1 year. Rats were inspected once or twice daily, mortality was recorded and rats in extremis were sacrificed and necropsied (see below). Ophthalmoscopic examination of both eyes was carried out before the start of the study and after 4, 14, 26, 52 and 79 weeks of treatment for 20 males and 20 females from each control group; the test groups were examined in a similar manner.

Surviving males were examined ophthalmoscopic ally at 110 weeks and 112/113 weeks (termination) and surviving females at 110 and 118/119 (termination). Food consumption was recorded weekly and water consumption was recorded over 3-day periods during each of the first four weeks, and during weeks 13, 26, 41, 52, 65, 78, 92 and 101. Body weight was recorded weekly for the first 12 weeks, then fortnightly until week 68, and then weekly until termination. Before the start of treatment and after 14, 26, 40, 53, 66, 79, 92 and 102 weeks, blood was taken for measurement of haematological and clinical chemistry parameters, and additionally in males at 111/112 weeks and in females at 118/119 weeks. Urine samples were collected periodically for urine analysis and for analysis of paraquat in the urine. Five animals of each sex per group were sacrificed at 52 weeks for estimation of concentrations of paraquat in the liver, lungs, kidneys, skin, plasma and urine. Necropsy was performed on all decedents, the 10 animals of each sex per group sacrificed at 52 weeks and those surviving to termination, and selected organs were weighed. These and other selected organs were preserved and processed for histopathological examination. Mortality was not affected by treatment and survival to termination was 38-55% in males and 47-50% in females. No clinical adverse effect was seen, except corneal opacity, which was seen at 150 mg/kg in males and at 75 mg/kg in females. At ophthalmoscopy, cataracts were seen at 150 mg/kg in both sexes and, from 103 weeks, at 75 mg/kg in both sexes. In the males, the prevalence of cataracts was not unequivocally increased at 25 mg/kg; however, a statistical analysis of the eye changes revealed an increase in posterior capsular changes at week 110 in the males at 25 mg/kg. Food intake at 150 mg/kg was reduced in both sexes, in the males for the first year of the study and in the females during the first 6 weeks; these changes were small. Depression of body-weight gain was seen at 150 mg/kg in both sexes, but was more severe in males and was also present in males at 75 mg/kg. Body-weight gain in males at 25 mg/kg and in females at 75 and 25 mg/kg was not different from that in the controls. Test material-related effects were not seen on haematological or clinical chemistry parameters, or on urine analysis. At 52 weeks, the concentration of paraquat in the urine was dose-related. In rats sacrificed at 52 weeks, paraquat was detected in the plasma and kidneys at all dietary concentrations, while paraquat was present in the lungs of animals at 75 and 150 mg/kg; only at 150 mg/kg and in females was paraquat found in the liver. Paraquat was found in some skin samples taken from males at 75 mg/kg and from both sexes at 150 mg/kg. No test material-related effects were seen on organ weights, other than those attributable to changes in body weight. Macroscopically, there was an increase in corneal opacity and focal sub-pleural changes at 75 and 150 mg/kg. Proliferative alveolar changes were also seen at these dietary concentrations. Lung histopathology was examined by two groups (Tables 4-6). An initial assessment of lung histopathology was made based on that of Life Sciences Research's own staff pathologists and of two consultant pathologists from the USA (Table 4). The other assessment was by Ishmael (Table 5), at that time Head of Pathology at Imperial Chemical Industries. There were some clear differences. Finally, the slides were examined by four independent pathologists and the results were reported by Busey (1986) (Table 6). It was concluded that the differentiation of bronchiolar adenomas and carcinomas from the non-neoplastic lesions typical of paraquat was difficult. However it was also concluded that the incidence of lung neoplasms in the test groups was comparable to that in the control groups (Ishmael & Godley, 1983; Woolsgrove, 1983; Sotheran et al., 1981; Woolsgrove, 1985; Busey, 1986; Ishmael, 1987).

Table 4. Initial assessment of lung histopathology in rats given diets containing paraquat

From Woolsgrove (1983)
^a p < 0.001
^b Bronchiolar-alveolar or squamous cell carcinomas

Table 5. Second assessment of lung histopathology in rats given diets containing paraquat

From Ishmael & Godley (1983)
^a p < 0.01

Table 6. Final assessment of lung histopathology in rats given diets containing paraquat

From Busey (1986)
^a p < 0.01

It was concluded from the data summarized in Table 6 that there was no association between the incidence of adenomas, carcinomas or the two combined, and exposure to paraquat. In contrast, there was a significantly increased incidence of adenomatosis at 150 mg/kg when all animals were included in the analysis (i.e. those sacrificed at 52 weeks, decedents and those killed at termination).

Ishmael (1987) reviewed the slides of the head region, in which squamous cell carcinomas of the skin and subcutis had been reported. In males, 11 such tumours were seen in the study (1, 2, 2, 0 and 6 in the two control groups and at the lowest, intermediate and highest doses, respectively) as originally reported and in Ishmael (1987). The site of origin of these tumours, however, differed and Ishmael (1987) suggested they should not be considered as a single phenomenon for statistical purposes. Other changes seen included dilatation of the fourth ventricle of the brain (hydrocephalus) in females at 75 and 150 mg/kg. Cysts and cystic spaces were seen in the spinal cords and, in males, prevalence was significantly greater than that in the controls in all test groups, although there was no clear dose-response relationship. This pathological change was found in females, but the frequency in test groups and control groups was similar (and similar to the frequency in the males in test groups). Degeneration of the sciatic nerve was found in males at 75 and 150 mg/kg. Changes were present in the eyes. At the highest doses, peripheral lenticular degeneration, more severe in females, and pear-shaped posterior peripheral lenticular change was seen. Mid-zonal lenticular degeneration, lens capsular fibrosis and/or lens ruptures were all seen. At 75 mg/kg, changes were milder. These changes were seen in both decedents and those rats surviving to termination. At the highest concentration, in the decedents, peripheral retinal degeneration was observed in females and proteinaceous vitreous humour was seen in males. Some changes were seen at the lowest dietary concentration; in male survivors these were moderate peripheral morgagnian corpuscles, slight peripheral lenticular degeneration, moderate mid-zonal lenticular degeneration and loss of outer nuclear layer of the retina. The last was unlikely to be a compound-related effect as the prevalence was lower in both the controls and at higher doses. In female survivors to termination, at the lowest dietary concentration, changes observed were moderate peripheral morgagnian corpuscles, slight peripheral lenticular degeneration and moderate mid-zonal lenticular degeneration.

At termination, there was no clear evidence of an effect on the retina at the lowest dose, although in males at the two higher dietary concentrations there may have been an effect on the periphery of the retina. This study was continued for a longer duration than that recommended by the OECD (104 weeks for long-term studies in rats). The NOAEL was 25 mg/kg for lenticular lesions after 103 weeks at 25 mg/kg in males and likewise in females at 103 and 110 weeks (see Table 7 for ophthalmoscopy findings at 103 weeks and Table 8 for lens findings at necropsy). This NOAEL is equivalent to 1.25 mg of paraquat ion/kg bw per day. This interpretation is supported by the findings from the other long-term studies in rats.

Table 7. Frequency of effects on the lens (in life) at 103 weeks in rats given diets containing paraquat

From Ishmael (1987)
* Greater incidence than combined control groups, statistically significant at p = 0.05 or less

Table 8. Frequency of effects on the lens at necropsy in rats given diets containing paraquat (all animals, regardless of time of death)

From Ishmael (1987)
* Greater incidence than combined control groups, statistically significant at p = 0.05 or less

Mice

In a 104-week study, groups of 80 male and 80 female JCL:ICR mice were given diets containing paraquat (paraquat dichloride; purity, 98%) at a dietary concentration of 0, 2, 10, 30 or 100 mg/kg, providing intakes of paraquat dichloride equal to 0, 0.26, 1.31, 3.92 and 13.09 mg/kg bw per day in males, and 0, 0.26, 1.32, 3.82 and 13.03 mg/kg bw per day in females. These intakes are equal to 0, 0.19, 0.95, 2.84 and 9.48 mg of paraquat ion/kg bw per day in males, and 0, 0.19, 0.96, 2.77 and 9.43 mg of paraquat ion/kg bw per day in females. At weeks 26 and 52, 10 male and 10 female per group were sacrificed. The mice were examined twice daily and adverse clinical effects, including mortality, were noted. Body weight and food consumption were measured weekly until week 26, and fortnightly thereafter. Blood was taken for haematology and clinical chemistry (including determination of plasma, erythrocyte and brain cholinesterase activities) from the animals killed at 26 and 52 weeks and from those that survived to termination. Urine analysis was performed on animals killed at 26, 52 and 104 weeks. Survivors were sacrificed at 104 weeks. Necropsy was carried out on the animals killed at 26 and 52 weeks and