Asia Pacific J Clin Nutr (1995) 4: 259-264

Diet and oral cancer - a case control study

MPR. Prasad*, TP Krishna*, S Pasricha*, MA Quereshi**, K Krishnaswamy*

*National Institute of Nutrition, Hyderabad; **MNJ Cancer Hospital, Hyderabad

---

Apart from strong genotoxic carcinogens, other environmental factors are implicated in both causes and prevention of cancers. A hospital based case control study was conducted to examine the role of diet in the aetiology of oral and oropharyngeal cancers. In this article, past dietary intake and nutrient estimates, obtained through diet history method and biochemical nutritional status at the onset of the disease are presented. The results of the study suggest that poor dietary intake of vegetables and fruits coupled with low estimated intake of betacarotene, thiamine, riboflavin, folate, vitamin C, iron and copper, modify the risk potential. The biological indicators of the nutritional status such as plasma vitamin A, E, red cell folate and plasma zinc were significantly reduced in cases and yielded moderate risk estimates. The risk estimates though of moderate magnitude are of importance in relatively homogeneous subjects with respect to diet and nutrition.

The findings are in line with several other epidemiological observations. The combined effects of micro nutrients appears to be protective in countering the adverse effects of exogenous exposures to tobacco. The protective role of vegetables and fruits is of potential interest in terms of etiologic causes and prevention.

---

Introduction

Oral cancer is one of the ten most common cancers in the world¹. In India it accounts for a third of all cancers². It is documented in all parts of India. Tobacco smoke and betel chewing with or without tobacco are suggested as etiologic factors accounting for 90% of all oral cancers^l. The mortality and morbidity of oral cancers are well documented in India. In fact tobacco related cancers seem to be a leading cause of cancers in the whole of the Indian sub-continent². The mortality of oral cancers despite advances in treatment and management appear to be high with poor chances of survival.

Though environmental exposure to tobacco and other carcinogens are important determinants of cancer, it is now well documented that other factors such as diet have an important role to play in the multistage carcinogenesis process. Several dietary factors act as risk modifiers^3-7. In general though, dietary deficiencies have not been shown to initiate events, epidemiological and experimental studies provide strong evidence for dietary substance, in promotion, progression and inhibition of cancer^8,9.

World wide epidemiological data suggest a strong protective role for fruits and vegetables^l0,ll. The protective effects of these in addition to non nutrients have been attributed to micronutrients s 1000 uch as vitamin A, b carotene, folate, vitamin E and vitamin C¹². Limited observation suggest that probably zinc and selenium deficiencies are also causally related^l3-14.

The surveys conducted by NNMB in 10 states of India document poor consumption of green leafy vegetables and also poor intake of vitamins A, C, folate and riboflavin in population groups¹⁵. As early as 1933, Orr in his classical description of oral cancer in India attributed vitamin A deficiency as one of the important contributing agents¹⁶. As there is meagre information on the dietary and nutritional correlates of oral cancer in India, a case control study of oral and oropharyngeal cancers was envisaged. The objectives of the study were to :

1. quantitate past dietary intake of food groups and nutrients
2. estimate the nutritional status at the onset of the disease and arrive at risk estimates.

Materials and methods

Cases :

The cases were patients of oral cancer, who presented themselves for the first time at the Government Cancer Hospital (Mehdi Nawab Jung) Hyderabad. Advanced cases and those who received treatment elsewhere were deleted. In all, 45 subjects of both sexes were included after excluding those who had previously consumed vitamin or mineral pills for several days.

Controls:

The controls were carefully matched for age, sex, socioeconomic status and habits and were drawn from attendants of the cancer patients or from patients suffering from non cancerous or other minor ailments.

A detailed clinical, anthropometric and personal history was taken and 10ml blood sample collected in an heparanised tube for biochemical analysis. A dietary history, using standard dietary cups were obtained from compliant individuals participating in the study.

Dietary details:

Dietary consumption pattern over the last six months was assessed by an oral questionnaire. The questionnaire was designed to obtain information regarding the dietary habits. Average frequency of consumption of food items in the preceding six months or before the change in the dietary habit and quantities of different food groups were obtained. From the frequency of consumption, number of food items consumed per month were obtained.

Biochemical Analysis:

Haemoglobin and albumin were estimated by cyanmethaemoglobin and dye binding methods respectively^17,18. RBC folate was estimated micro-biologically using L. Casei¹⁹. Riboflavin and thiamin levels were estimated indirectly by measuring the erythrocyte glutathione reductase (EGR) and erythrocyte transketolase (ETK) activity and expressed as activity coefficients^20,21. High pressure liquid chromatography was used for plasma vitamin A, vitamin E was determined by thin layer chromatography^22,23. Serum iron was estimated colorimetrically while the trace metals, copper, magnesium and zinc were estimated by atomic absorption spectrophotometry^24,25.

Statistical Analysis:

The Mann-Whitney test was applied for comparisons of means between the two groups²⁶. Odds ratios (ORs) were computed (using control values as cut off points) after the intake of nutrients were adjusted for total calorie intake^27,28. Wherever significant odds ratios were observed, 95% confidence intervals (CI) were also included.

Results

The 1000 mean age in our study was 54 years. More than 90% were either tobacco smokers or tobacco chewers. Smoking was a consistent feature. None of the patients had the habit of reverse smoking. The distribution of cancers by sites are given in Table 1. The body weights of patients varied between 31.0 to 67.0kgs and a mean weight of 45.2± 1.29kgs and the mean body mass index (BMI) was 0.17± 0.003. The controls had a mean weight of 49.3± 1.38kgs and mean BMI of 0.19± 0.005 as shown in Table 2.

Table 1. Age, weight, site of cancer, and TNM classification   Cases Controls Age (years) 53.9± 1.7 53.5± 1.8 Weight (kg) 45.2± 1.29 49.3± 1.38 Site of cancer a) Tongue 18   b) Oral cavity 19   c) Oropharynx 8   Number of cases by TNM classification T1N0M0 8   T1N1M0 T2N0M0 10   T2N1M0 T2N2M0 20   T2N1M1 T2N2M1 2   T3N1M0 T3N3M0 5   Values Mean ± SD. No. of subjects in each group = 45

Table 2. Blood nutrients and other parameters Parameter Case Control Body mass index 0.17± 0.003** 0.19± 0.005 Haemoglobin g/ l 120.5± 1.97 123.2± 1.85 Albumin g/ l 32.3± 0. 1000 45** 34.6± 0.40 RBC folate nmol/ l 277.4± 4.79* 295.3± 6.54 EGR (AC) 1.27± 0.019 1.25± 0.018 ETK (AC) 1.22± 0.019 1.20± 0.02 Vit A m g/ dl 27.3± 1.9** 39.6± 1.66 Vit E m g/ dl 570.0± 34.5*** 920.6± 57.97 Serum iron m mol/ l 16.5± 0.94 16.9± 0.47 Copper m mol/ l 19.9± 0.35 19.9± 0.50 Magnesium m mol/ l 1.4± 0.037** 1.2± 0.033 Zinc m mol/ l 9.8± 0.21*** 14.2± 0.58 Values are mean ± SE. No. of observations 45 in each group * p<0.02 **p<0.005 ***p<0.001

The protein energy status as reflected by serum albumin was significantly lower than that of controls (Table 2). More than 60% of the cases were in first tertile (Fig. 1). The haemoglobin levels and the serum iron status were not different between cases and controls.

Figure 1. Tertile distribution of percent cases and controls for serum albumin. (p<0.05) and RBC folates (p< 0.05)

Figure 2. Tertile distribution of percent cases and controls for plasma vitamin A (p<0.01) and vitamin E (p< 0.001).

However, of the B-complex vitamins, red cell folate was significantly lower in cases (p <0.02). The status of other B complex vitamins, viz. thiamin and riboflavin were not affected.

Among fat soluble vitamins both vitamin A (p <0.005) and vitamin E (p <0.001) were significantly low. The tertile distribution indicated in Figure 2 reflects the same.

The two trace metals magnesium and zinc were significantly different from the controls with the former being higher while the latter was lower in cases, with all the case subjects falling in the first tertile (Figure 3). Serum copper level was not affected.

Figure 3. Tertile distribution of percent cases and controls for plasma magnesium (p <0.02) and zinc (p <0.001).

The odds ratios in Table 3 indicated that the risk was higher for low haemo 1000 globin, albumin, folate and vitamin A; whereas the risk was higher for high magnesium levels. Plasma zinc and vitamin E values were not presented in the table as the populations of the two groups were well apart giving an infinite odds ratio. Dietary assessment Food Groups The major food groups such as cereals, pulses, milk, vegetables and fruits were quantitated (Table 4). The vegetable intake green or otherwise was significantly low as compared to controls. In fact, the intake of both these food groups was only a third to half of that observed in controls. When the frequencies of several food items (Table 5) were analysed, more frequent intake of other vegetables were associated with lower risk of oral cancer (p <0.01), while the OR indicated a higher risk with lower consumption of vegetables and higher consumption of flesh and fried foods. Nutrients The nutrient intakes were computed as per the Nutritive Value of Indian Foods from the diet history. It was obvious that the cases and the controls were taking many fewer calories (Table 6)29. When all the nutrients were adjusted to the energy intake, the intake of carotene, riboflavin and magnesium were significantly higher in controls. The odds ratios show higher risk associated with low intakes of carotene, thiamin, riboflavin, folic acid, vitamin C, iron, magnesium and copper. Discussion Epidemiological observations throughout the world clearly demonstrate that environmental exposure to tobacco is an important factor in the aetiology of oral cancerl. Literary evidences document not only an association between tobacco and oral cancers but also highlight the risk modifying properties of certain specific food items such as vegetables and fruits and also of nutrients such as vitamin A, C and betacarotene30-32. Macro and micro nutrient deficiencies or imbalances have long been considered as probable causes or risk modifiers of cancer of the oral cavity33. Our study aimed to investigate the influence of dietary habits on the risk of tobacco induced oral and oropharyngeal cancers which were diagnosed within 2 months of the onset of signs and symptoms. It provided an opportunity to study the role of diet in the aetiology of oral and oropharyngeal cancers. The tumour node metastasis (TNM) classification, suggest that most of them were diagnosed in the early stages of tumour development with or without the involvement of local lymph nodes with a few exceptions. The findings of the present study either on the dietary intake prior to the onset of the disease or the nutrient profile at the onset, may be considered as indicators of natural sequence in the evolution of disease process, as not all nutrients are affected but only those that are aetiologically related. The general nutritional status of the cases as reflected by body weight, BMI and albumin, appeared to be lower than the controls. We have earlier documented that protein energy malnourished individuals were at greater risk for developing cancers as they have a higher metabolic susceptibility to carcinogen such as polycyclic aromatic hydrocarbons34-37. Th 1000 e nutrient profiles estimated at the onset of the disease, indicated that vitamin A, E, zinc and folate-deficient subjects were susceptible to cancers. Several investigators have reported the protective effects offered by vitamins A and C rich foods38-40. The fact that in the present study the estimates of prior dietary intake indicated low intakes of betacarotene, riboflavin and vitamin C rich foods, favours the causal relationship of the nutrient deficiencies to the development of the disease process. The relative risk estimates, though of moderate magnitude, further reinforce the aetiological relationships as the subjects investigated were relatively homogeneous with respect to dietary intake.

Table 3. Odds ratios (ORs) with 95% confidence intervals (CI) for oral cancer by tertile by nutrients in blood Nutrient OR CI Haemoglobin 1.00 2.57 2.86 (0.82-8.08) (0.92-8.90) Albumin 1.00 1.44 4.33 (0.41-5.05) (1.36-13.82) RBC Folates 1.00 3.92 3.80 (1.16-13.240) (1.11-13.030) EGR 1.00 0.76 1.45 (0.27-2.15) (0.53-3.96) ETK 1.00 1.90 1.32 (0.67-5.40) (0.48-3.59) Vitamin A 1.00 1.38 4.62 (0.36-5.35) (1.47-14.52) Iron 1.00 0.27 0.87 (0.08-0.95) (0.31-2.42) Copper 1.00 0.42 0.87 (0.14-1.23) (0.31-2.43) Magnesium* 1.00 1.73 6.00 (0.47-6.38) (1.77-20.31) *The comparison in this is done with 1st tertile value. No. of observations = 45 Table 4. Average intake (g) of food groups Foods groups Cases (16) Controls (16) Cereals 259± 33.0 313± 32.7 Pulses 23.4± 4.34 24.8± 4.56 Green leafy vegetables 9.09± 0.97 26.29± 8.08* Other vegetables 50.1± 11.79 106.4± 19.45** Milk & milk products 84.3± 16.23 70.8± 18.06 Flesh foods 15.4± 6.08 10.3± 3.18 Poultry products 8.1± 6.12 7.5± 2.39 Fruits 24.4± 18.09 53.9± 13.56 Oils/ fats 15.6± 2.66 13.9± 1.89 Figures in parenthesis indicate no. of observations. All values are mean ± SE *P <0.05 **P <0.02

Table 5. Mean frequency of intake of food groups/ month with odds ratios (ORs) and 95% confidence intervals (CI) for oral c 1000 ancer

1000

Foods	Cases*	Controls*	ORa	CI
Cereals	75.2± 7.8	86.3± 3.8	0.43
Pulses	26.8± 5.3	31.9± 6.6	1.00
Green leafy vegetables	5.93± 2.0	6.9± 2.0	1.80	0.39-8.21
Other vegetables, roots & tubers	20.7± 3.6	46.6± 6.2**	¥
Milk and milk products	9.4± 3.5	6.9± 3.8	0.51
Animal foods	4.9± 1.9	4.2± 0.99	2.60	0.52-13.04
Fruits	7.0± 2.2	7.5± 2.2	1.29
Fried foods	16.2± 5.2	3.3± 2.0	--
Fish	0.75± 0.28	0.19± 0.13	0.18
Tea/ coffee	61.9± 9.6	37.5± 7.5	0.20
Fats and oils	0.6± 0.24	0.5± 0.31	0.21
Spices	34.5± 5.8	31.5± 6.3	0.89

No. of subjects in each group = 16. a: Median of the control taken as the cut-off point. *Mean± SE ** p< 0.01

Table 6. Calorie adjusted mean nutrient intake and odds ratio (ORs) with 95% confidence intervals (CI) for oral cancer

Nutrients	Casesb	Controlsb	ORa	CI
Calories	1938.21± 0.34	1984.27± 0.34	--	--
Proteins (g)	49.85± 0.30	47.20± 0.28	0.45	--
Fat (g)	35.55± 0.42	28.56± 0.42	0.45	--
Calcium (mg)	545.00± 0.57	596.90± 0.45	1.29
Carotenea (mg)	330.67± 0.50	655.59± 0.56*	3.00	0.67-13.40
Thiamin (mg)	0.68± 0.43	0.81± 0.35	2.20	0.52-9.30
Riboflavin (mg)	0.66± 0.38	1.74± 1.48*	4.33	0.88-21.30
Folic acid (m g)	103.75± 0.40	128.65± 0.36	4.33	0.88-21.30
Vitamin C (mg)	39.66± 0.52	43.68± 0.59	2.20	0.52-9.30
Iron (mg)	21.26± 0.30	22.49± 0.30	1.67	0.41-6.82
Magnesium (mg)	217.22± 0.40	327.49± 0.33**	5.57	1.13-27.52
Copper (mg)	3.05± 0.43	3.91± 0.32	4.33	0.88-21.31
Zinc (m g)	5308.8± 0.51	5766.3± 0.30	0.78

No. of observations in each group = 16; a: negligible amount of vitamin A present (not included in carotene); b:Mean± SE; *p< 0.05; **p< 0.01

 

It is interesting to note that dietary sources for betacarotene and folate are similar, namely vegetables, of which intake appears to be low. There is evidence to show that a precancerous condition has been associated with low folate intake⁴¹. Population based registries in India document a higher incidence of cervical cancers⁴². It is possible that folate in general may be involved in differentiation of epithelial cells as cancers at the above sites were all squamous cell carcinomas. Our observation on oesophageal cancer indicated low red cell folate levels. Infants in India are born with inadequate stores of folate and several other nutrients and continue to subsist on inadequate intakes of micro nutrients^42,43.

The sources of vitamin E in the Indian diet are vegetable oils and cereals. Though the intakes of these two items were not significantly different, the intakes in general, in both cases and controls were much less than the recommended dietary allowances. It is possible that in cases where other nutrient deficiencies exist, the prooxidant and antioxidant balance is tilted towards adverse effect. Current research evidences support a strong preventive role for all antioxidants mentioned above in the causation of cancers at different sites. It was therefore not surprising to observe that vitamins A, C, E and betacarotene appeared to be protective in our series of cases. In the multistage carcinogenic process, which is complex, antioxidants appear to function as an inhibitor of promotional aspects of cancer⁴⁴.

In our earlier study on oesophageal cancers, results indicated a greater risk in individuals with low plasma zinc in addition to low vitamin A levels⁴⁵. The discovery of the role of zinc in cancer is recent. Plasma zinc is known to decrease in several cancers¹³. Studies in China, Africa, Iran and Russia show a positive association between dietary zinc deficiency and higher incidence of oesophageal carcinomas^46-51. Our study is one of the first to highlight an association between plasma zinc and oral cancer. The Indian diets which are cereal and pulse based are higher in phytates which can decrease absorption of zinc and contribute towards a zinc deficient state⁵². The intake of zinc particularly in the low income group is half of the recommended dietary allowance for Indians²⁹.

The experimental findings in relation to chemical carcinogenesis support our present observations of higher risk for zinc-deficient subjects particularly for nitrosamine induced cancers⁵³. Zinc is known to have a significant effect on glutathione-S transferase and also on DNA repair mechanisms⁵⁴.

The relation of oral cancer with magnesium is intriguing. We have no satisfactory explanation for higher plasma magnesium levels in cases as compared to controls. The other non-nutrient dietary items examined were in line with the observations of Notani and Jayant on upper aerodigestive cancers wherein more frequent co 1000 nsumption of tea/ coffee, chillies and less frequent intake of fish appeared to increase the risk⁵⁵.

We are aware of the limitation in interpretation of our results. Familiarity with the hypothesis is a potent source of bias. Though the relationship between dietary constituents (nutrients and non nutrients) and cancer is complex, the consistent observation across wide population groups in different countries strongly support vitamins A, C and betacarotene as candidates for chemoprevention agents^56,57. Our findings must be assessed in this context as the results are in agreement with previous epidemiological work in several countries^7,30,38.

Thus the micronutrients in the diet appear to play a pivotal role in reducing damage due to environmental exposures and probably act synergistically to enhance several protective mechanisms against carcinogenesis.

Acknowledgements

We thank Dr. Vinodini Reddy, Director, National Institute of Nutrition for her keen interest and valuable suggestions. We thank Mr KP Dalvi, Mr VK Goud, Mr PN Rao for their excellent technical assistance.

---

Diet and oral cancer - a case control study

MPR. Prasad, TP Krishna, S Pasricha, MA Quereshi, K Krishnaswamy

Asia Pacific Journal of Clinical Nutrition (1995) Volume 4, Number 2: 259-264

---

References

1. World Health Organization, Control of oral cancer in developing countries-- a WHO Meeting. Bulletin of WHO Geneva 1984; 62: 817-830.
2. Annual Report, Population cancer registries-- New Delhi: Indian Council of Medical Research, 1989.
3. Micozzi MS. Foods, micronutrients, and reduction of human cancer. In: Moon TE, Micozzi MS, eds. Nutrition and cancer prevention. New York Marcel Dekker Inc, 1989: 213-242.
4. Lippman SM, Kessler JF, Meyskens FL Jr. Retinoids as preventive and therapeutic anticancer agents (Part I). Cancer Treat Rep 1987; 71: 391-405.
5. Lippman SM, Kessler JF, Meyskens FL Jr. Retinoids as preventive and therapeutic anticancer agents (Part II). Cancer Treat Rep 1987; 71: 493-515.
6. Tuyns AJ, Riboli E, Doornbos G. Nutrition and cancer of the oesophagus. In: Jossens JV, Hill MJ, Geboers J, eds. Diet and human carcinogenesis, Amsterdam Elsevier, 1985; 71-79.
7. Marshall J, Graham S, Mettlin C, Shedd D, Swanson M. Diet in the epidemiology of oral cancer. Nutr Cancer 1982; 3: 145-149.
8. Newberne PM, Schrager TF, Canner MW. Experimental evidence on the nutritional prevention of cancer. In: Moon TE, Micozzi MS, eds. Nutrition and cancer prevention. New York Marcel Dekker Inc, 1989: 33-82.
9. Hunter DJ, Willet WC. Human epidemiological evidence on the nutritional prevention of cancer. In: Moon TE, Micozzi MS, eds. Nutrition and cancer prevention. New York Marcel Dekker Inc, 1989: 83-100.
10. Steinmetz KA, Potter JD. Vegetables, fruit and cancer. I, Epidemiology. Cancer causes and control, 1991; 2: 325-327.
11. Ziegler RG. Vegetables, fruits and carotenoids and the risk of cancer. Am J Clin Nutr 1991; 53: 251s-259