Guidelines for Canadian Drinking Water Quality: Guideline Technical Document – Trifluralin

Exposure

In 1987, trifluralin was detected in several municipal water supplies in Saskatchewan at trace (nanogram per litre) levels.^{Footnote 12} It was not detected in drinking water supplies of 77 municipalities in Manitoba or Alberta (detection limits 0.05 to 0.5 µg/L).^{Footnote 13 Footnote 14} Trifluralin was detected in one of 91 wells at 41 µg/L in a 1984 survey in southern Ontario.^{Footnote 15} Trifluralin has occasionally been detected at trace levels (below 1 µg/L) in surface waters in Manitoba.^{Footnote 13} Trifluralin was not detected (detection limit 0.1 µg/L) in an eight-week sampling of irrigation water in southern Saskatchewan.^{Footnote 16} Based on a concentration of 0.05 µg/L (or half the usual detection limit of 0.1 µg/L), the estimated median Canadian exposure is 0.08µg/d, or 1× 10^-6 mg/kg bw per day from drinking water.

In Canada, trifluralin was not detected in a national survey of 120 foods (detection limit 4 ppb).^{Footnote 17} In the United States, trifluralin was not detected in over 27 000 food samples covering 27 crops (detection limit 10 ppb).^{Footnote 2 Footnote 18} The theoretical maximum dietary intake of trifluralin is estimated to be 0.0271 mg/d, or 0.00039 mg/kg bw per day for an adult, based on the assumption that the maximum permitted residues of 0.1 mg/kg are present in all wheat, peas, beans, tomatoes and turnips consumed.^{Footnote 19 Footnote 20} Actual residues and intakes are expected to be much lower than this estimate.

Analytical Methods and Treatment Technology

Trifluralin may be monitored in water by extraction with a solvent such as dichloromethane and quantification by gas chromatography with electron capture detection. Confirmatory identification is by gas chromatography-mass spectrometry.^{Footnote 16 Footnote 21} Using this method, the detection limit was 0.05 µg/L in a Manitoba survey^{Footnote 13} and 0.1µg/L in a Saskatchewan survey.^{Footnote 16} The U.S. EPA reported that trifluralin is removed with 100% effectiveness by conventional treatment using alum, sedimentation and filtration. In addition, it may be removed from drinking water by reverse osmosis, granular activated carbon and air stripping.^{Footnote 21}

Health Effects

Absorption of trifluralin from the gastrointestinal tract is dose-dependent, with about 20% absorption in rats given an oral dose of 100 mg/kg bw^{Footnote 22} and almost 90% absorption with an oral dose of 1 mg/kg bw.^{Footnote 23} After administration of an oral dose of 1 mg/kg bw to rats, 87% was eliminated within 60 hours, 47% via faeces and 40% via urine. Most of the trifluralin was metabolized by several pathways to 17 metabolites, whereas about 10% was eliminated unchanged in the faeces. Only 2.1% remained in rat tissues after 96 hours, with the highest concentration (0.22 ppm) in liver and lower concentrations (0.1 ppm or less) in kidney, fat, lung, spleen, intestine and stomach.^{Footnote 23} The acute oral toxicity of trifluralin is low. The primary effects of chronic ingestion include reduced body weights, increased liver weights and manifestations of renal toxicity, including glomerulonephritis and renal calculi.

In a 1984 study, four groups of beagle dogs (six per sex per group) were fed 0, 30, 150 or 750 ppm trifluralin in the diet, equivalent to 0, 1, 4.8 and 25 mg/kg bw per day, for one year. Histopathological examination revealed no systemic or organotoxic effects at any dose. Diarrhoea, vomiting, decreases in body weight gain and increases in liver and spleen weights, along with changes in serum chemistry (lipid and protein fractions), were noted at the highest dose.^{Footnote 21 Footnote 24 Footnote 25} The no-observed-adverse-effect level (NOAEL) in this study was 4.8 mg/kg bw per day.^{Footnote 24 Footnote 25}Trifluralin has been tested in six long-term (two-year) dietary studies, four on rats and two on mice. In the first two bioassays, carried out in the early 1960s on Harlan and on Cox rats, no treatment-related tumorigenic effects were observed. These tests were not adequate by today's standards because of methodological limitations and contamination problems.^{Footnote 2 Footnote 21} In a National Cancer Institute (NCI) study on Osborne-Mendel rats and on B6C3F₁ mice in which technical trifluralin (later shown to be contaminated with 84 to 86 ppm dipropylnitrosamine [NDPA]) was administered in the diet, there were no tumorigenic effects in rats of either sex or in male mice, but hepatocellular carcinomas and alveolar/bronchiolar adenomas were observed in female mice.^{Footnote 21 Footnote 26} However, this could have been related to nitrosamine contamination of the trifluralin, as many nitrosamines have been demonstrated to be tumorigenic in mice.^{Footnote 26} In a 1980 study undertaken by the registrant, purified trifluralin with NDPA contamination below 0.01 ppm was administered in the diet to F344 rats (60 per sex per group) and B6C3F₁ mice (80 per sex per group). Non-tumorigenic effects in both sexes of rats included reduced kidney, heart, thymus, pancreas and body weights, elevated blood urea nitrogen, hypertrophy of the liver and testes, glomerulonephrosis and formation of renal calculi; the NOAEL was 813 ppm, or approximately 41 mg/kg bw. Effects were similar but less marked in mice. There was no evidence of oncogenicity in female rats or in either sex of mice, in contrast to the results of the NCI study. In male rats, significant dose-related increases in benign and malignant tumours of the urinary transitional epithelium of the kidney and bladder were observed. The combined incidences of benign and malignant tumours in male rats were 0/60, 3/60, 6/60 and 7/60 for doses of 0, 41, 163 and 315 mg/kg bw.^{Footnote 18 Footnote 21} The registrant claimed that the renal tumours were specific to F344 rats that were subject to chronic renal disease, glomerulonephrosis and renal calculi. In a subsequent 90-day dietary study on males of this strain, minor changes in kidney function, reversible after a six-week recovery period, were demonstrated.^{Footnote 24 Footnote 25} The World Health Organization (WHO) agreed with the registrant's position that oncogenic effects of trifluralin were not demonstrated in the bioassay.^{Footnote 23} Trifluralin was not mutagenic with or without metabolic activation in a number of microbial tests on various organisms, including Ames tests on Salmonella typhimurium.^{Footnote 22 Footnote 27 Footnote 28} Tests were also negative in a dominant lethal test in Drosophila, an in vitro mouse lymphoma system^{Footnote 21} and several mammalian in vivo tests, including a dominant lethal assay and a chromosomal aberration study in mice (unpublished studies cited in WHO background document prepared for reference 24). An in vitro test on human lymphocyte cultures was positive for chromosomal aberrations,^{Footnote 29} as were several in vivo mouse bone marrow assays at high ^{Footnote 30 Footnote 31} but not low concentrations (1/100 of LD₅₀).

It is not clear whether or not the trifluralin used in these studies was significantly contaminated with nitrosamine.

Trifluralin did not cause teratogenic effects in rats, dogs or rabbits at doses up to 1000 mg/kg bw.^{Footnote 2 Footnote 21 Footnote 24} Manifestations of embryotoxicity, including cardiomegaly, wavy ribs, body weight reductions and increases in percentages of runts, occurred at maternally toxic doses above 10 and 20 mg/kg bw in rats and dogs, respectively.^{Footnote 2 Footnote 21 Footnote 24} Reproductive effects, including decreased litter size and increases in spontaneous abortions and stillbirths, were observed in rabbits above 225 mg/kg bw by gavage on days 6 to 18 of gestation^{Footnote 21 Footnote 24} and above 200 ppm in the diet (10 mg/kg bw) in two-, three- and four-generation reproductive studies in rats.^{Footnote 5 Footnote 11 Footnote 21 Footnote 24 Footnote 25}

Rationale

Based on evaluation by the Food Directorate of the Department of National Health and Welfare, a negligible daily intake (NDI) was established as follows:^{Footnote 25}

NDI = [4.8 mg/kg bw per day] / 1000 = 0.0048 mg/kg bw per day

where:

• 4.8 mg/kg bw per day is the NOAEL observed in a one-year feeding study in dogs in which changes in liver and spleen weights and in serum chemistry (lipid and protein fractions) were noted^{Footnote 25}
• 1000 is the uncertainty factor (×10 for interspecies variation; ×10 for intraspecies variation; and ×10 for limitations of the data base).

An maximum acceptable concentration (MAC) was derived from this NDI as follows:

MAC = [0.0048 mg/kg bw per day × 70 kg bw × 0.20] / 1.5 L/d ≈ 0.045 mg/L

where:

• 0.0048 mg/kg bw per day is the NDI, as derived above
• 70 kg bw is the average body weight of an adult
• 0.20 is the proportion of total daily intake arbitrarily considered to be ingested in drinking water (the theoretical maximum residues in Canadian food are at present less than 10% of the NDI; apportionment of 80% of intake to other sources allows for additional intake from food, should tolerances be set in future)
• 1.5 L/d is the average daily consumption of drinking water for an adult.

The MAC is based on a non-carcinogenic endpoint. However, because dose-related increases in tumours of the transitional cell epithelium of the kidney and bladder have been observed in one sex of one species of rodent after dietary administration of trifluralin in a single study,^{Footnote 18} and because the structurally similar compound ethylfluralin is considered to be carcinogenic in animals, the potential lifetime cancer risks associated with exposure to trifluralin have been estimated^{Footnote 32} to ensure that the MAC is protective.

The maximum potential lifetime cancer risk associated with the ingestion of 1 µg/L trifluralin in drinking water has been estimated by the robust linear extrapolation method, employing an interspecies correction for body weights, to be 4.5 × 10^-7, based on the combined incidence of benign and malignant tumours of the kidney and urinary bladder. The lifetime risk from drinking water containing trifluralin at the guideline value of 45 µg/L is therefore estimated to be 2.0 × 10^-5, slightly above a range that is considered to be "essentially negligible."

References