Administration of long-term tamoxifen therapy modifies the plasma lipoprotein-lipid concentration and lipid transfer protein I activity in postmenopausal women with breast cancer.

Tamoxifen remains one of the most effective agents in the treatment of breast cancer. However, the development of persistent side effects from chronic administration of tamoxifen remains a concern. The objective of this project was to investigate the effect of long-term tamoxifen therapy (2 years) on the plasma lipoprotein concentration and lipid transfer activity in postmenopausal women diagnosed with breast cancer. Two distinct populations of breast cancer patients were recruited for this study: postmenopausal women diagnosed with breast cancer that have never been on tamoxifen and postmenopausal women diagnosed with breast cancer that have been on tamoxifen (20 mg once daily) for 2 years (n = 18 each group). Blood was collected for total and lipoprotein cholesterol (C) and triglyceride (TG) analysis by established enzymatic assays prior to and 2 years following the initial tamoxifen dose. To determine the effect of tamoxifen administration on lipid transfer between lipoprotein fractions, lipid transfer protein (LTP I) activity was measured. A significant decrease in total and low-density lipoprotein (LDL) cholesterol levels and a moderate increase in high-density lipoprotein (HDL) triglyceride levels were observed in plasma samples from postmenopausal women with breast cancer who were administered tamoxifen for 2 years. No significant differences in total and lipoprotein C and TG plasma levels were observed in samples from women with breast cancer that never received tamoxifen. LTP I activity was significantly decreased in patients receiving tamoxifen compared to patients who never received tamoxifen. These specific tamoxifen-induced effects may be important for a number of reasons. Although the apparent decrease in total C and LDL-C levels are favorable for reducing the risk of cardiovascular disease, the elevated levels of HDL-TG have been related to the increased risk of ischemic heart disease. Furthermore, understanding how tamoxifen influences LTP I activity provides valuable insight into how the administration of tamoxifen modifies plasma lipoprotein-lipid levels.

Micronucleus frequencies in urothelial cells of catheterized patients with chronic bladder inflammation.

Epidemiological studies suggest an association between chronic inflammation and increased risk for cancer, although the mechanism underlying this relationship is unresolved. In the present study, we test the hypothesis that DNA damage is induced in the epithelium of tissues during such inflammation by products of activated inflammatory cells. Individuals on long-term indwelling urinary catheterization were used as a study population. These individuals have chronic bladder inflammation and, as a population, an increased risk for bladder cancer. Urine of 29 patients and 26 age-matched non-catheterized controls was collected and micronucleus (MN) frequencies were determined in exfoliated urothelial cells in the urinary sediments. The urine from the catheterized group had large numbers of white blood cells (mean count, 26.6 +/- 3.6 cells per high-power field), indicating the presence of a chronic bladder infection and an inflammatory reaction. In contrast, white blood cells were not present in urine from individuals in the control group. There was no significant difference in MN frequencies in the 2 groups (mean frequencies, controls: 0.098 +/- 0.030%; catheterized: 0.140 +/- 0.025%, p = 0.13). These data imply that chromosomal damage does not always occur during chronic inflammation. Although the reasons for this observation are yet to be determined, possible explanations include the pathophysiology of the inflammatory reaction and the influence of vitamins, non-steroidal anti-inflammatory drugs and the catheter itself in protection against inflammatory cell-mediated DNA damage.

DNA adducts in bronchial biopsies.

To investigate the feasibility of measuring DNA-carcinogen adducts in the lungs of non-surgical patients, endobronchial biopsies were obtained from 78 patients undergoing routine diagnostic bronchoscopy. Lung cancer was present in 37 (47%) of the patients. DNA was isolated from the tissues and analyzed by HPLC- or nuclease-PI-enriched 32P-postlabelling, using procedures selective for aromatic adducts. Chromatograms from all 28 current smokers showed a distinctive diagonal adduct zone which was present in only 24 of 40 ex-smokers and 4 of 10 lifetime non-smokers. Adduct levels and chromatographic patterns were similar in bronchial tissue from different lobes of the lung, in bronchial and alveolar tissue, and in tumor and non-tumor bronchial tissue taken from the same subject. Bronchial DNA adduct levels were strongly associated with cigarette smoking status and dropped rapidly after smoking ceased. Higher levels of DNA adducts seen in the lung-cancer patients were mainly due to cigarette smoking. Frequent alcohol intake was the only dietary factor associated with higher levels of bronchial DNA adducts. We conclude that the level of bronchial DNA adducts is strongly associated with cigarette-smoking history and with alcohol intake, but is not associated with lung cancer independently from its relation to smoking. The results indicate the feasibility of using 32P-postlabelling to detect and quantitate genetic damage in bronchial biopsy specimens.

Carcinogen adducts as an indicator for the public health risks of consuming carcinogen-exposed fish and shellfish.

A large variety of environmental carcinogens are metabolically activated to electrophilic metabolites that can bind to nucleic acids and protein, forming covalent adducts. The formation of DNA-carcinogen adducts is thought to be a necessary step in the action of most carcinogens. Recently, a variety of new fluorescence, immunochemical, and radioactive-postlabeling procedures have been developed that allow the sensitive measurement of DNA-carcinogen adducts in organisms exposed to environmental carcinogens. In some cases, similar procedures have been developed for protein-carcinogen adducts. In an organism with active metabolic systems for a given carcinogen, adducts are generally much longer lived than the carcinogens that formed them. Thus, the detection of DNA- or protein-carcinogen adducts in aquatic foodstuffs can act as an indicator of prior carcinogen exposure. The presence of DNA adducts would, in addition, suggest a mutagenic/carcinogenic risk to the aquatic organism itself. Vertebrate fish are characterized by high levels of carcinogen metabolism, low body burdens of carcinogen, the formation of carcinogen-macromolecule adducts, and the occurrence of pollution-related tumors. Shellfish, on the other hand, have low levels of carcinogen metabolism, high body burdens of carcinogen, and have little or no evidence of carcinogen-macromolecule adducts or tumors. The consumption of carcinogen adducts in aquatic foodstuffs is unlikely to represent a human health hazard. There are no metabolic pathways by which protein-carcinogen or DNA-carcinogen adducts could reform carcinogens. Incorporation via salvage pathways of preformed nucleoside-carcinogen adducts from foodstuffs into newly synthesized human DNA is theoretically possible.(ABSTRACT TRUNCATED AT 250 WORDS)

Clastogenic activity in urine of workers occupationally exposed to pesticides.

Genotoxicity in the urine of orchardists occupationally exposed to pesticides was investigated. Urine samples were obtained during pre-spraying and spraying periods from 22 non-smoking orchardists who spray large amounts of pesticides during the fruit growing season. For comparison purposes, urine was collected from 11 non-smoking personnel at an agricultural research station located near the application site and from 21 non-smoking individuals (reference controls) in a non-agricultural area. Organic material was isolated from urine by preparative reversed-phase high-pressure liquid chromatography, and assayed for clastogenic activity using Chinese hamster ovary cells. Urine samples collected during the pre-spraying period showed no significant differences in clastogenic activity compared to that found for the reference control group. However, clastogenic activity of urine specimens collected during the spraying period was significantly elevated (p less than 0.001) for the highly-exposed orchardists, but not for the research station personnel. Clastogenicity of orchardists' urine was observed within 8 h of pesticide application.

32P-postlabeling detection of DNA adducts in fish from chemically contaminated waterways.

Fish were collected from sites in the chemically-contaminated Buffalo River, New York, and the Detroit River, Michigan. The sediments of these rivers have high levels of chemical contaminants, including polycyclic aromatic hydrocarbons (PAHs), and fish from these locations have high prevalences of liver cancer. To determine chemical-DNA interactions and a possible role for chemicals as a cause of the observed tumors, DNA was isolated from livers and was enzymatically digested to normal and adducted nucleotides. The DNA digests were enriched for hydrophobic, bulky adducts, either by preparative reverse phase high pressure liquid chromatography, or by selective nuclease P1 dephosphorylation of normal nucleotides. DNA-chemical adducts were then quantitated by 32P-postlabeling analysis. Regardless of the adduct enrichment procedure, the chromatograms derived from DNA of fish from polluted areas showed a diffuse, diagonal radioactive zone consisting, at least in part, of multiple overlapping discrete adduct spots. The behavior of the adducts in the diagonal radioactive zone and of their unlabeled precursors is consistent with their identification as nucleotide adducts of a variety of bulky, hydrophobic, aromatic genotoxic compounds. Analysis of bile demonstrated recent exposure to multi-ringed aromatic compounds.

Carcinogenic and genotoxic activity of extracts from contaminated sediments in western Lake Ontario.

In order to investigate the putative association between chemical contamination in western Lake Ontario and high prevalences of fish tumors, sediments from Hamilton Harbour and Oakville Creek in Lake Ontario and reference sites in non-polluted areas of Ontario, Canada were collected and extracted for organic contaminants. Sediment extracts from Hamilton Harbour had the highest concentrations of polychlorinated biphenyls and organochlorine insecticides (ppb) and contained very high concentrations of polynuclear aromatic hydrocarbons (ppm); although the levels of these compounds varied widely with sampling location in the harbor. A sediment extract from Hamilton Harbour was mutagenic in the Ames bacterial assay, both with and without microsomal activation. High levels of aromatic DNA adducts were induced in cultured mouse C3H1OT1/2 cells after in vitro exposure to Hamilton Harbour sediment extract. In two separate carcinogenicity experiments involving a sac fry microinjection assay with rainbow trout (Oncorhynchus mykiss), Hamilton Harbour sediment extract induced hepatocellular carcinomas in fish. No hepatic neoplasms were observed in fish that had been treated with sediment extract from Oakville Creek, or with extract from a reference sediment. The significance of these results is discussed in relation to the distribution of neoplasms in feral fish within western Lake Ontario.

HPLC enrichment of hydrophobic DNA--carcinogen adducts for enhanced sensitivity of 32P-postlabeling analysis.

The sensitivity of 32P-postlabeling analysis for the quantitation of DNA-carcinogen adducts can be sharply increased by enriching adducted nucleotides relative to normal nucleotides in digested DNA samples prior to postlabeling. Normal and adducted nucleotides from benzo[a]pyrene-adducted DNA were injected onto a reverse phase HPLC column. Normal nucleotides were eluted with 5% methanol, 95% 1 M ammonium formate, pH 3.5. Hydrophobic adducted nucleotides were then recovered from the column using either a linear gradient of methanol (to fractionate adducts) or a step gradient of methanol to elute all hydrophobic adducts in one fraction. Recovered chromatographic fractions were dried, postlabeled with 32P, and aromatic DNA--carcinogen adducts were detected and quantitated using TLC on polyethyleneimine cellulose. Results from the HPLC adduct enrichment procedure were generally similar to those from an alternative enrichment scheme utilizing nuclease P1 to selectively dephosphorylate normal nucleotides. HPLC and nuclease P1 procedures are easily combined for the dual analysis of DNA samples from organisms exposed to environmental carcinogens.

DNA adducts, micronuclei and leukoplakias as intermediate endpoints in intervention trials.

Internal dosimeters that can provide information about responses to chemopreventive agents in a short time would be invaluable for planning treatment protocols for large-scale intervention trials. Micronuclei meet many of the prerequisites of a good intermediate endpoint. They can be quantified in cultured cells, animal tissues and human exfoliated cells and biopsies. With image scanning, up to 10(5) cells can be screened for micronuclei within a few minutes. The predictive value of micronuclei has been demonstrated using cultured cells exposed to carcinogens and chemopreventive agents and using oral mucosa of betel-quid chewers. DNA adducts, as detected by 32P-postlabelling techniques, could conceivably be another potentially useful marker. However, prior to their use in intervention trials, interindividual variations in their levels in primary, secondary and nontarget tissues and the relationship with doses of carcinogens must be established. The wide scatter of DNA adduct levels in the bronchial mucosa of smokers and of nonsmokers reveals one difficulty that can be encountered using this marker in intervention trials.

32P-postlabeling analysis of aromatic DNA adducts in fish from polluted areas.

Brown bullheads (Ictalurus nebulosus) were sampled from sites in the Buffalo and Detroit Rivers where fish are exposed to high levels of sediment bound polycyclic aromatic hydrocarbons, and suffer from an elevated frequency of liver cancer. DNA was isolated from the livers of these wild fish and from control specimens which were raised in clean aquariums. DNA was enzymatically digested to normal and adducted nucleotides, and hydrophobic/bulky adducts were enriched in the digests either by preparative reverse-phase high-pressure liquid chromatography, or selective nuclease P1 dephosphorylation of normal nucleotides. Aromatic DNA-carcinogen adducts were then quantitated using 32P-postlabeling analysis. Using both adduct enrichment procedures, chromatograms derived from DNA of fish from polluted areas showed a diffuse diagonal radioactive zone not present in DNA from aquarium raised fish. The diagonal zone appeared to consist at least in part of multiple overlapping discrete adduct spots which could be partially separated by gradient high-pressure liquid chromatography prior to 32P-postlabeling analysis, and most of which were more strongly retained on a reverse-phase column than the major benzo(a)pyrene-DNA adduct. The behavior of the adducts in the diagonal radioactive zone and of their unlabeled precursors is consistent with their identification as nucleotide adducts of a variety of bulky hydrophobic aromatic environmental compounds. Total pollution-related adduct levels as analyzed by HPLC adduct enrichment and 32P-postlabeling were 70.1 +/- 29 (SD) nmol/mol normal nucleotide in fish from the Buffalo River, and 52 and 56 nmol/mol for two specimens from the Detroit River.

Use of intermediate endpoints in quantitating the response of precancerous lesions to chemopreventive agents.

A current area of emphasis in cancer research is the determination of whether cancer can be prevented through the use of naturally occurring chemopreventive agents such as beta-carotene. A major area of concern in the design of long-term, large-scale population studies to ascertain the efficacy of such chemopreventive agents lies in the paucity of biological data on the activity of these agents in man. The studies described in this paper were performed to determine whether a series of short-term markers could be used in chemopreventive trials as indicators of the possible success of a chemopreventive regime. Three such markers are described. The first involves the measurement of genotoxic damage in the target tissues of carcinogen-exposed individuals by using the micronucleus test on exfoliated cells. This end point has been successfully used to demonstrate a reduction in carcinogen damage (micronuclei production) in the oral cavity of individuals in population groups at elevated risk for oral cancer (tobacco and betel quid users in the Philippines, snuff users in the Northwest Territories). The second marker involves the determination of DNA adducts in exfoliated cells of carcinogen-exposed individuals by the use of DNA postlabelling procedure. The final marker discussed involves a chemical determination of the levels of a chemopreventive agent in target tissues of individuals receiving a supplement in the diet. In this case, the example described is beta-carotene in exfoliated cells of carcinogen-exposed individuals. These three markers may be combined to determine whether a chemopreventive agent reaches a target tissue, affects DNA adduct formation, and prevents genotoxic damage.(ABSTRACT TRUNCATED AT 250 WORDS)

Relationship between cellular levels of beta-carotene and sensitivity to genotoxic agents.

The usefulness of an in vitro test system to predict the inhibitory effect of beta-carotene on the genotoxic activity of carcinogens/mutagens was explored. To facilitate the comparison of data obtained from cultured cells (CHO) and from exfoliated human cells, endpoints were used which can be quantitated in both cell systems: the frequency of micronuclei for estimating the effect of genotoxic agents, and cellular levels of beta-carotene as a protective agent. In CHO cells, beta-carotene inhibited the clastogenic and micronucleus-forming effect of methyl methanesulfonate (MMS) and 4-nitroquinoline-1-oxide (4NQO), but had no protective action against gallic acid, tannic acid, and aqueous extract of areca nut or H2O2. The extent of inhibition depended on the ratio of beta-carotene to MMS. Doses of beta-carotene which exerted a protective effect in vitro ranged from approximately 2 to 5 ng per 10(6) CHO cells. Comparable levels of beta-carotene were previously found to reduce the frequency of micronucleated exfoliated cells from the buccal mucosa of tobacco and areca-nut chewers (Stich et al., 1984b).

32P-postlabelling analysis of aromatic DNA adducts in human oral mucosal cells.

Exfoliated mucosal cells were collected from the oral cavity of three groups at high risk for oral cancer: Indian betel nut chewers, Filipino inverted smokers (burning end of cigar in mouth) and Indian Khaini tobacco chewers. DNA was extracted from these samples, as well as from samples of exfoliated cells of Canadian non-smoking controls. DNA was analyzed for the presence of aromatic DNA adducts using 32P-postlabelling analysis. Five chromatographically distinct adducts were found in samples from both the high risk groups and the nonsmoking controls. Individual adducts were detectable in approximately 30-95% of samples, depending on the adduct and population group. Estimated levels of specific adducts ranged from non-detectable (prevalence relative to normal nucleotides less than 1 X 10(-9)) to occasionally greater than 1 X 10(-7). No adducts were found in high risk groups which did not also appear in control subjects.

Beta-carotene levels in exfoliated mucosa cells of population groups at low and elevated risk for oral cancer.

Beta-carotene was estimated in exfoliated oral mucosa cells in groups of individuals at various risks for oral cancer. Approximately 4 X 10(6) exfoliated cells were collected from each subject by brushing the oral mucosa. Cell pellets were hydrolyzed with pronase and then with KOH/methanol. Beta-carotene was extracted with hexane, separated by reverse-phase HPLC, and detected at 450 nm. Mean beta-carotene levels in exfoliated cells were 0.08 ng/10(6) cells for 56 heavy consumers of alcoholic beverages (150 g or more per week), 1.36 ng/10(6) cells for 28 Seventh Day Adventists (all abstainers from alcohol, tobacco and meat consumption), 1.39 ng/10(6) cells for 55 lacto-vegetarians of the International Society for Krishna Consciousness (ISKC) (abstainers from alcohol and tobacco), and 1.08 ng/10(6) cells for 61 representatives of a "Western" life-style pattern (64% consumed the equivalent of at least one bottle of wine or 7 bottles of beer per week, and all were non-smokers). If the heavy alcohol consumers (males) are matched to non-drinking males of comparable age, the mean beta-carotene values are 0.08 ng versus 1.24 ng/10(6) cells. The possible involvement of the low levels of beta-carotene in the mucosa of heavy alcohol drinkers in increased sensitivity towards the carcinogenic and genotoxic activity of cigarette smoking plus alcohol ingestion is discussed.

Beta-carotene levels in exfoliated human mucosa cells following its oral administration.

Beta-carotene levels of exfoliated oral mucosa cells can be increased severalfold by the oral administration of this provitamin. Beta-carotene was estimated by HPLC analysis in pronase-treated exfoliated cells obtained by brushing the entire oral mucosa with a toothbrush. A small percentage of individuals did not respond with an increase of beta-carotene in their mucosa cell in spite of a relatively large intake of the provitamin (360 mg in 4 days, or 2880 mg in 16 weeks, respectively). Levels of beta-carotene in the mucosa cells are affected by the concurrent administration of vitamin A: 0.27 ng beta-carotene per 10(6) cells in the placebo group, 1.79 ng following the intake of beta-carotene (180 mg/week for 16 weeks), and 4.29 ng after beta-carotene (180 mg/week for 16 weeks) plus vitamin A (100,000 IU/week for 16 weeks) consumption. The considerable variations in tissue levels of beta-carotene following its oral administration must be taken into account when cancer intervention trials using this agent are designed and evaluated.

A pilot beta-carotene intervention trial with Inuits using smokeless tobacco.

The frequency of exfoliated cells with micronuclei (MNC) was used to estimate the genotoxic effect of smokeless tobacco (snuff) on the oral mucosa and to follow the response to the administration of beta-carotene (180 mg/week, given twice weekly in 6 capsules of 30 mg each). The pilot trial was carried out with Inuits in Gjoa Haven, Northwest Territories, Canada. Their traditional diet, which is rich in caribou and seal meat and liver but low in vegetables and fruits, leads to "normal" serum levels of retinol (447 ng/ml in non-users of tobacco and 463 ng/ml in tobacco users) but low levels of beta-carotene (57 ng/ml for non-users of tobacco and 47 ng/ml for users). Prior to the twice-weekly administration of beta-carotene, the frequency of MNC was 1.87% +/- 0.92 (n = 23) in the mucosa of the lower gingival groove where the tobacco was usually kept. It decreased significantly (P less than 0.001) to 0.74% +/- 0.42 following the 10-week oral administration of beta-carotene capsules. The frequency of MNC did not change significantly in the group receiving a placebo and in snuff users who received no treatment over the 10-week trial period. The size and morphological appearance of the typical snuff-related, whitish, wrinkled patches of the mucosa where the tobacco was kept was not affected by the 10-week treatment with beta-carotene. Similarly, no reduction was observed in the frequency of anucleated, exfoliated mucosa cells. Beta-carotene appears to be an efficient inhibitor of MNC in the oral mucosa of snuff users who do not suffer from any vitamin A deficiency and who have "normal" levels of retinol.

High-pressure liquid chromatography for isolation of clastogenic agents from urine.

Small columns of XAD-2 resin have been widely used to extract and concentrate mutagenic materials from urine. Using analytical HPLC and assays for clastogenicity with Chinese hamster ovary cells, we found that small columns of XAD-2 resin (1.5 ml bed volume) retain only a small percentage of organic material and undetectable amounts of genotoxic activity in urine samples. Increasing the size of the XAD resin bed resulted in better recoveries, but much organic material was still lost by overloading of the column. In contrast, when urine was acidified and chromatographed by preparative reversed-phase HPLC using large-bed-volume (500 ml) commercial columns, retention of hydrophobic organic material from urine was excellent. Subsequent stepwise elution of the column with increasing concentrations of acetone produce 3 fractions of organic material of increasing hydrophobicity. When urine from smokers was analysed, all 3 fractions contained material which was clastogenic to Chinese hamster ovary cells. The procedures developed are suggested as a new general purpose approach to the isolation of genotoxic materials from urine.

Clastogenic agents in the urine of coffee drinkers and cigarette smokers.

Organic material from the urine of smokers, coffee drinkers, and controls was extracted and separated into 3 fractions of differing hydrophobicity using preparative reversed-phase high-pressure liquid chromatography. Fractions were assayed for clastogenic activity using Chinese hamster ovary cells. Smoking, coffee drinking, or both habits together resulted in a substantial increase in the genotoxicity of organic material in all 3 fractions. The clastogenicity of fractions 1 and 2 (the two most hydrophilic) was abolished by the addition of either catalase or superoxide dismutase to the Chinese hamster ovary cell system, suggesting the involvement of active oxygen species in the clastogenic response. Clastogenicity of fraction 3, however, was resistant to the action of catalase and superoxide dismutase. Fractions were tested for their ability to generate hydrogen peroxide in vitro during a 10-min incubation at elevated pH. Fractions 2 and 3, but not fraction 1 from smokers, coffee drinkers, or those with both habits generated significantly more hydrogen peroxide at high pH than did the corresponding fractions from control subjects. For fractions 2 and 3 but not 1, the ability of a sample to generate hydrogen peroxide at high pH was positively correlated with its ability to generate chromosome aberrations at neutral pH in tissue culture. The data indicate that both coffee drinking and cigarette smoking result in the appearance of clastogenic materials in urine, and suggest that such clastogenic agents may produce chromosome aberrations via the production of active oxygen species.

Determination of free and protein-bound N-nitrosoproline in nitrite-cured meat products.

Levels of nitrosoproline (NPRO) in the urine of subjects on a Western-style free-choice diet typically ranged from 1 to 2 ng/ml. Higher levels of NPRO in the urine were associated with the consumption of nitrite-cured meat products. The elevation in NPRO excretion after the consumption of cured meats was not diminished by the concurrent consumption of the nitrosation inhibitor ascorbic acid, indicating the presence of preformed NPRO or some NPRO precursor in the meat. Cured meat samples were analysed for NPRO using methanolic extraction, followed by gas chromatography and nitrosamine quantification by means of a thermal energy analyzer. Levels of NPRO detected in meats were less than those indicated by the amount of NPRO excreted when the same meat was eaten. Digestion of meat samples by proteolytic enzymes sharply increased the amount of NPRO detectable by chemical analysis. Nitrite-cured meat products contained 39-3900 ng total NPRO/g product. Of this amount of NPRO, 3-93% was protein-associated and only detectable after proteolytic digestion. Protein-associated NPRO may represent nitrosated N-terminal proline in proteins or peptides. Such NPRO would not be detectable as free NPRO on direct chemical analysis, but would be released by proteolysis in the digestive tract when the meat is eaten.

The effect of dietary factors on nitrosoproline levels in human urine.

The effect of dietary components on the levels of nitrosoproline ( NPRO ) excreted over a 24 h period in the urine was examined in volunteers ingesting known amounts of various food products. The ingestion of nitrite-preserved meats (85-170 g per meal), including canned, rolled or Yunnan ham, cured pork, luncheon meat, and various Chinese and European-style sausages, led to urinary NPRO excretion levels ranging from 2.5 to 78.5 micrograms/24 h, whereas the consumption of non-preserved meat and fish products, including chicken, herring, salmon, shrimp, ground beef (hamburger), pork chops and beef liver, led to relatively low NPRO excretion levels, ranging from 0.0 to 0.8 micrograms/24 h. The urinary NPRO levels of 22 vegetarians and 14 lacto-vegetarians averaged 0.8 and 1.4 micrograms/24 h, respectively. A change from a nitrite-preserved meat diet to a vegetarian diet was accompanied by an approximately six-fold reduction in urinary NPRO levels; however, these remained above control levels for at least 3 days following the dietary change. The relatively high NPRO levels following the ingestion of nitrite-preserved meats could not be reduced by nitrite-trapping chemicals, including ascorbic acid, ferulic acid, caffeic acid, or phenolic-containing mixtures such as coffee and tea, which were effective in suppressing endogenous NPRO formation following the intake of nitrate and proline. The high urinary NPRO levels after ingestion of preserved meat products appear to be due to the consumption of preformed NPRO . An understanding of the relative contribution of preformed and endogenously formed nitrosamines appears to be essential when designing dietary intervention programmes.