International Agency for Research on Cancer (IARC) - Summaries & Evaluations


VOL.: 37 (1985) (p. 37)

5. Summary of Data Reported and Evaluation

5.1 Exposure data

Smokeless-tobacco habits are practised by many millions of people, principally in Africa, Asia, Europe and North America, utilizing several techniques, products and dosage levels. In some countries, average consumption by users is estimated to be about 5 kg per year.

Among the thousands of compounds present in tobacco, the tobacco-specific nitrosamines are the only identified carcinogens that occur in mg/kg concentrations. Low levels (mg/kg) of carcinogenic polynuclear aromatic hydrocarbons and metals and of the a-emitting 210Po (0.1-1.0 pCi/g) have also been detected. Use of smokeless tobacco entails extensive exposure to relatively high levels of tobacco-specific nitrosamines.

5.2 Experimental data

Various chewing tobaccos and unburnt cigarette tobaccos and their extracts were tested by oral administration in mice, by topical application to the oral mucosa of mice, rats and hamsters, and by subcutaneous administration, skin application, inhalation, intravesicular implantation and intravaginal application to mice. All of these studies suffered from certain deficiencies.

In a two-stage, mouse-skin assay, applications of tobacco extract followed by promotion by croton oil induced papillomas and squamous-cell carcinomas of the skin. In further two-stage, mouse-skin assays, application of tobacco extracts following initiation by 7,12-dimethylbenz[a]anthracene resulted in papillomas.

A commercial Swedish snuff was tested for carcinogenicity in rats, by topical administration in a surgically-created oral canal, alone or in combination with herpes simplex virus type 1 infection. Two squamous-cell carcinomas of the oral cavity were observed in the group receiving both treatments, but this result was not statistically significant.

Snuff was tested by oral administration in hamsters, alone and in combination with calcium hydroxide, but the data were insufficient for evaluation. Several studies in hamsters in which snuff was administered as single or repeated applications into the cheek pouch or fed in the diet yielded insufficient data for evaluation.

Subcutaneous injection of ethanol extracts of snuff to rats did not produce an increase in tumour incidence.

Nass was tested for carcinogenicity in hamsters by administration into the cheek pouch or by skin application. No tumour was found at the site of application. Although nass was associated with an apparent excess of liver tumours in various groups receiving cheek-pouch administrations, which may be indicative of carcinogenic activity, deficiencies in reporting do not allow an evaluation to be made.

Ethanol extracts of chewing tobacco (Nicotiana tabacum) induce mutations in Salmonella typhimurium and in Chinese hamster V79 cells. They also induce micronuclei in bone-marrow cells of Swiss mice.

Ethyl acetate extracts of a chewing tobacco induce sister chromatid exhanges in cultured human lymphocytes and in a human lymphoblastoid cell line. Ethyl acetate and ethanol extracts of this tobacco induce transformation in Syrian hamster embryo cells.

Aqueous extracts of nass and khaini induce chromosomal aberrations in Chinese hamster ovary cells.

Saliva collected during the chewing of an Indian tobacco induce chromosomal aberrations in Chinese hamster ovary cells.

An increased proportion of micronucleated cells was found in exfoliated oral-mucosa cells from users of khaini and nass.

Sister chromatid exchanges are induced in Chinese hamster ovary cells by anatabine, nicotine and nornicotine.

5.3 Human data

Oral leukoplakia, a precancerous lesion, has been associated with oral-snuff use in a number of studies. One study of shammah users and several studies of nass users showed the same association.

Epidemiological studies of cancer and the oral use of smokeless tobacco in western populations have often not distinguished between tobacco chewing and snuff usage. Studies that have are summarized first.

Chewing tobacco

Reports of series of oral-cancer patients indicate that a high proportion were tobacco chewers and that the cancer often developed at the site at which the quid was placed habitually. However, data on chewing tobacco often came only from medical records; coexistent smoking habits often were not mentioned.

In two of five case-control studies in which data on tobacco use were appropriately obtained, the proportion of tobacco chewers among patients with cancer of the oral cavity, pharynx or larynx was two to three times higher than in control subjects; however, confounding by tobacco smoking or alcohol consumption could not be excluded. A large study of oral, pharyngeal and oesophageal cancer reported no difference in chewing-tobacco use between cases and controls; although the relative risk of having cancer of the oral cavity or pharynx was increased in tobacco chewers, this study is not convincing because of major discrepancies in the tabulated data. Data on dose-response are lacking in all three studies. The other two case-control studies provide no clear evidence that tobacco chewing is associated with oral cancer: one study was very large but did not control for smoking, and one had serious methodological limitations.

Results from the four case-control studies of chewing-tobacco use and cancer of the oesophagus tend to show a slight increase in incidence. Nose and nasal-sinus cancers were found to be unrelated to tobacco chewing in one case-control study. No association between chewing tobacco and bladder cancer was observed in five case-control studies.

No cohort study of chewing tobacco alone and cancer has been reported.

Oral snuff

Reports of case series indicate that a high proportion of oral-cancer patients took snuff orally, and that the cancer frequently developed at the site of snuff application.

Four case-control studies, three from the south-eastern USA and one from Scandinavia, have implicated snuff use in the etiology of cancer of the oral cavity and, to a lesser extent, of the pharynx. In three of these studies, relative risks could not be computed; however, the differences in snuff usage between cases and controls were substantial, and confounding by cigarette smoking could be largely excluded. In the fourth study, in the south-eastern USA, the relative risk of oral and pharyngeal cancer for white women who used snuff but did not smoke was four times that for women with no tobacco habit; a strong dose-response relationship was observed; adjustment for other risk factors did not substantially reduce the relative risks.

In a cohort study of snuff users with non-malignant oral lesions, none developed cancer; however, the study was inadequately reported, had methodological limitations, and therefore could not be satisfactorily interpreted.

One case-control study has suggested that oral use of snuff may be associated with certain types of nasal-sinus cancer; in other case-control studies, no association was evident between snuff use and bladder cancer or between snuff use and cancer of the oesophagus.

Smokeless tobacco, unspecified

Studies that have not distinguished snuff from chewing tobacco are informative for four reasons when considered in conjunction with the habit-specific studies summarized above. First, reports of three case series confirm the high relative frequency of smokeless-tobacco use in oral-cancer patients. Four case-control studies have reported smokeless-tobacco use to be moderately to strongly associated with oral cancer, although smoking habits were not controlled for in three of the studies.

Second, a dose-response relationship was found in one large case-control study. The relative risks for oral cancer in men, after adjustment for other risk factors, ranged from four-fold for moderate smokeless-tobacco use to more than six-fold for heavy use.

Third, two cohort mortality studies, in which large numbers of persons with and without unspecified smokeless-tobacco habits were followed, provide evidence of a positive association with cancer. There was a two- to three-fold increased risk of death from oral, pharyngeal and oesophageal cancer in one study and from oesophageal cancer in the second.

Fourth, studies of unspecified smokeless-tobacco use provide some evidence of an increased risk of cancers at sites outside the upper digestive and respiratory tracts.

Whereas the data summmarized above all come from studies in North America and western Europe, the data below refer to studies of oral use of tobacco and nasal use of snuff in South-East Asia and in Africa.


Oral cancer in users of mishri and gudakhu has been studied only in prevalence surveys; no case was found.


Oral cancers were seen in users of shammah.

Tobacco plus lime (khaini)

Two large case control-studies, from Pakistan and India, reported two-fold to 14-fold increases in the risk of oral-cancer occurrence in tobacco (presumably tobacco-lime) users relative to non-users, in smokers and nonsmokers considered separately. Indirect evidence, deducible from various other studies of chewing and oral cancer in which the predominant habit entailed use of tobacco and lime without areca nut, corroborates the existence of this increased cancer risk.

Tobacco plus lime plus other components

In two case series, the majority of oral-cancer patients used nass; in another, the cancers were found to develop at the site at which the quid was placed habitually. Two case-control studies showed five-fold to 20-fold increases in the risk of oral cancer in association with nass use in the USSR; however, adjustment was not made for smoking habits and other potential confounders.

Use of naswar, examined in one case-control study in Pakistan, was associated with a marked increase in oral-cancer risk; however, positive confounding by tobacco smoking and betel-quid chewing could not be eliminated.

Nasal snuff

Two case-control studies among Bantu subpopulations in Africa, among whom nasal and oral use of indigenous snuff (containing tobacco and other ingredients, including aloe) are common, showed a moderately elevated risk of nasal-sinus cancer in relation to this habit; however, the studies had severe methodological limitations.

In India, two studies (one cross-sectional, one prospective) of oral cancer found no association between oral cancer and snuff inhaling. A case-control study reported snuff inhaling to be more common among patients with cancers of the oesophagus, hypopharynx or oropharynx than among controls; however, adjustment was not made for other risk factors for these cancers.

No study was available that specifically addressed the possible carcinogenicity of nasal use of snuff formulated in North America or western Europe.

5.4 Evaluation

There is sufficient evidence that oral use of snuffs of the types commonly used in North America and western Europe is carcinogenic to humans. There is limited evidence that chewing tobacco of the types commonly used in these areas is carcinogenic.

Epidemiological studies that did not distinguish between chewing tobacco and snuff provide sufficient evidence for the carcinogenicity of oral use of smokeless-tobacco products, as reported in these studies.

In aggregate, there is sufficient evidence that oral use of smokeless tobacco of the above types is carcinogenic to humans.

There is sufficient evidence that oral use of tobacco mixed with lime (khaini) is carcinogenic to humans.

There is inadequate evidence that oral use of the other smokeless-tobacco preparations considered (nass, naswar, mishri, gudakhu and shammah) is carcinogenic to humans.

There is inadequate evidence that nasal use of snuff is carcinogenic to humans.

There is inadequate evidence to evaluate the carcinogenicity of chewing tobacco, snuff or nass to experimental animals.

For definition of the italicized terms, see Preamble Evaluation.

Subsequent evaluation: Suppl. 7 (1987)

Last updated: 21 April 1998

    See Also:
       Toxicological Abbreviations