plant foods and cancer risk
TRANSCRIPT
Plant Foods And Cancer RiskScience And Tradition
John D. Potter
and fruitS,5. It presents, in abbreviated form, the available data on thequestion of whether the consumptionof vegetables and fruit is associatedwith a lower risk of cancer.
COHORT STUDIES
Particular vegetables and fruitshave long been believed to be usefulin the prevention and cure of disease. Western medicine, until quiterecently, largely involved the prescription of specific plants and foods. Inancient Egypt, then later in Greeceand Rome, a wide variety of useswere found for many different plantsand extracts1• Garlic was considereda holy plant in Egypt. Cruciferousvegetables were cultivated primarilyfor medicinal purposes and were usedtherapeutically against several acuteand chronic disorders2• Celery, cucumber, endive, and a variety of herbsand vegetable juices had standarduses. Peas were used in a conditionthat may have been angina pectoris.
The Romans believed that len
tils were a cure, for diarrhoea andconducive to an even temper (it is notclear whether these are related). Several fruits also were used therapeutically: the citron (similar to the lemon)was believed to act as an antidote forpoisons; raisins and grapes were incorporated into oral preparations,enemas, inhalations and topical applications3.
The medical traditions of Indiaand China, particularly', have retainedand extended much of this therapeutic knowledge. Today, in the developed world once again, there aremany interesting properties ascribedto edible plants, herbs, spices andfoods. Until relatively recently, however, both ancient and modern therapies and preventive strategies were
based on many generations of clinical experience and not on evidencederived from epidemiological and experimental studies.
DIET AND CANCER
In 1981, Doll and Peto produceda wide estimate of 10 to 70 per cent ofall cancer as being attributable todiet4• This conclusion was based onstudies showing increased risk inassociation with certain foods - particularly foods of animal origin. Overthe past decade, however, many studieshave examined the effect of plantfood consumption on health and disease. High consumption of vegetablesand fruit is protective against cancers at many sites; this conclusionappears to be better supported byliterature than almost all other dietaryhypotheses.
There are many biologically plausible reasons why consumption ofvegetables and fruit might deter theoccurrence of cancer. These includethe presence of potentially anticarcinogenic substances such as carotenoids, vitamin C, vitamin E, selenium, dietary fibre, dithiolthiones,isothiocyanates, indoles, phenols,protease inhibitors, allium compounds,plant sterols, limonene and others.However, the crucial issue is the complexity and variety of the mixtures,not the specific dose of one or twocompounds.
This paper presents an updateof our earlier review of the role ofplant foods - particularly vegetables
At least 21 cohort studies? haveexamined the relationship betweenvegetable and fruit consumption anddifferent types of cancer. Of these,19 found an inverse association for atleast one category of vegetable and/or fruit, and in 12 studies statisticallysignificant reductions in risk were reported. Of these studies, four reportedon all cancer sites; four on lung cancer; three each on colorectal and stomach cancers; two each on pancreatic, bladder and breast cancers; andone on prostate cancer.
It is lung cancer for which thecohort study evidence is the mostconsistent; inverse associations forvegetable and/or fruit consumptionhave been shown in populations ofNorwegian men, postmenopausal USwomen, Seventh Day Adventists, andmembers of a retirement communityin California. The association between
CONTENTS
•Plant Foods And Cancer Risk
- Science And Tradition
1
- John D. Potter
-
•Nutrition News 5
•Food Safety Evaluation -
National and InternationalPerspectives
6- Ramesh V. Bhatt and Pulkit Mathur•
Foundation News 8
colon cancer and vegetable consumption is not as consistent in the cohortstudies as in the case-control studiesreviewed below.
CASE-CONTROL STUDIES
Most of the evidence for an inverse association with vegetable andfruit consumption comes from casecontrol studies. As early as 1933,such a study by Stocks and Karns inBritain suggested such an association between intake of certain vegetables against cancer at all sites.Four hundred and sixty-two cancerpatients and 435 patients without cancer, each of whom provided a diethistory, were compared.
The most common cancers werethose of the breast, the colon andrectum, the uterus and the tongue.Cases reported lower consumptionof carrots, turnips, cauliflower, cabbage, onions, watercress and beetrootthan controls. Intake of these vegetables combined was statistically significantly lower in the cases than inthe controls. Although the methodsemployed in this study were crude bymodern standards, it is striking tonote that this apparently protectiveeffect of vegetable consumption andof milk, along with the risk-enhancingeffects of pipe smoking and beerdrinking were the major findings ofthis study.
Another study, in the same yearin India, although less rigorous indesign, again found that vegetableconsumption was higher in those whodid not develop oral cancer comparedwith those who did9•
Since that time, nearly 200 morecase-control studies worldwide haveexamined the relationship betweenvegetable and fruit consumption andcancer risk. Stomach cancer has beenthe most studied, followed by cancers of the colon, oesophagus, lung,oral cavity and pharynx, rectum andbreast; more than 10 studies havefocussed on each of these sites. Statistically significant inverse associations have been reported for one ormore vegetable and/or fruit categories in more than 70 per cent of thestudies for cancers of the followingsites: stomach, oesophagus, lung,oral cavity and pharynx, endometrium,pancreas, colon and skin. Prost~tecancer is the only cancer for whichthe majority of studies have not re-
ported at least one statistically significant inverse association; and indeed, for prostate cancer, no studyshows an inverse association. Someof these sites will now be reviewed inmore detail.
STOMACH CANCER
Thirty-one case-control studiesdealing with the relationship betweenvegetable and fruit consumption andstomach cancer have been reported.Twenty-eight of these studies foundan inverse association between cancer and consumption of one or morevegetable and/or fruit.
The most consistent and abundant findings appear to be for theconsumption of fruit, raw vegetables(including lettuce) and allium vegetables. Contrary to the pattern of inverse associations between intakesof the above vegetables and freshfruits and stomach cancer, severalstudies suggested that consumptionof potatoes and canned fruit werepositively associated with risk.
Potential confounders of the relationship between stomach cancerand vegetable and fruit consumptioninclude socioeconomic status. Lower
socioeconomic status (as measuredby education level, occupation, and/or income) was found to correlatewith stomach cancer risk in many ofthe studies. In addition, the study byCorrea et a/10 was conducted in Louisiana where stomach cancer ratesare much higher among blacks, whoare generally poor, as compared towhites in the area. Diet is likely to becorrelated with socioeconomic status and the consumption of fresh vegetables and fruits, which tend to berelatively expensive, may be muchlower among the lower socioeconomicsections of a population. Nonetheless, several of the odds ratios were
adjusted for the confounding eff~~tsof occupation, education, ethnlcltyand/or income as surrogates for socioeconomic status. These studiesstill noted a statistically significantinverse association between vegetableand fruit consumption and stomachcancer.
Consumption of pickled, fermented, salted and/or smoked foods,consumption of alcohol, consumption of large amounts of carbohydrate (as potatoes or rice), infrequentmeals, non-ownership of a refrigerator, poor dentition, residence in a
2
rural vs urban area (which may betied to socioeconomic status), history of chronic gastritis, blood type 0and smoking are other risk factors forstomach cancer elucidated by someof the studies reviewed here. Smoking is not as strong a risk factor forcancer of the stomach as for cancersofthe respiratory and upper gastrointestinal tracts.
COLON CANCER
Fifteen of the 21 case-controlstudies dealing with colon cancerfounda statistically significant inverse relationship between cancer and at leastone index of vegetable and fruit consumption. Among specific categoriesof vegetables, the results for cruciferous vegetables were the most striking in that most of the studies presenting findings for cruciferous vegetables found higher intake to be associated with lower risk.
Most of the studies showing nosignificant association or only positive associations with vegetable andfruit consumption were among theearliest conducted, which may be areflection of more crude dietary methodology or perhaps less attention paidto detail in gathering fruit and vegetable data (hypotheses regardingthese foods were not obvious at thetime these earlier studies were conducted).
OESOPHAGEAL CANCER
Fourteen case-control studies ofoesophageal cancer have examinedthe associations with vegetable andfruit consumption. Twelve found astatistically significant inverse relationship between intake of one ormore vegetables and/or fruits andoesophageal cancer.
Consumption of vegetables, specifically, was found to be invers~lyassociated in almost all of the studiesthat reported on this exposure and inmost cases, this was statistically significant. The association with potatoes appears to be the most inconsistent. Consumption of fruit in general.was inversely associated with oesophageal cancer.
Consumption of red chilli powder was also found to be a risk factor11• Although not a vegetable perse, such a risk may also apply towhole red chillies themselves. Alternatively, the risk may be derived from
some part of the drying process oruse of red chilli powder may be amarker for use of some other spicethat is the actual agent.
Smoking and heavy alcohol consumption are well established as therisk factors that account for the majority of oesophageal cancers in theindustrialised countries. These habits are associated with less healthfuleating patterns and poor nutritionalstatus. Thus, an obvious question iswhether the inverse associations foundwith vegetable and fruit consumptionin most of these case-control studiesare confounded by alcohol intake andsmoking. In fact, several of the studies reviewed here supply evidencethat vegetable and fruit consumptionhas an independent effect with manyof the odds ratios being adjusted forsmoking or alcohol consumption.Strong consistency is observed acrossstudies: adjustment for smoking andheavy drinking does not reduce risksassociated with diet.
Some areas of the world, suchas parts of Iran, South Africa andChina, experience very high rates ofoesophageal cancer that are not necessarily attributable to smoking andalcohol consumption. In further support of an independent effect of vegetable and fruit consumption, inverseassociations were found12 betweenintake of many vegetables and fruitsand oesophageal cancers in Iran although the study in Un Xian, China,was essentially null. In these areas,identified risk factors include low socioeconomic status, consumption of opiumand of known carcinogens in food,perhaps consumption of extremelyhot beverages and diets high in wheatand corn.
LUNG CANCER
Thirteen case-control studies oflung cancer have examined the effects of vegetable and fruit consumption. Eleven found lower risk associated with one or more vegetablesand/or fruits. Many of the studies foundconsumption of carrots (or dark yellow-orange vegetables) to be associated with lower risk and somewhatfewer reported a similar result for leafygreen vegetables (also described askale, dark green vegetables or spinach). Carrots and greens are two ofthe most concentrated dietary sourcesof a variety of carotenoids.
In these studies, the benefit con-
ferred by high intake of vegetableswas often more evident in smokersthan nonsmokers but not all studiesexamined associations within thesmoking strata. One explanation forthe trend for stronger associations insmokers might be that the majority ofcases in most studies of lung cancerare smokers and there may not havebeen enough power to detect trueassociations in nonsmokers. Anotherpossibility is that the benefit conferredby vegetable consumption is a latestage event that occurs after initiation by substances in cigarette smoke.Interestingly, a study of only nonsmoking female patients in Hong Kong(64 per cent of female lung cancerpatients in Hong Kong have neversmoked), also found inverse associations for intakes of vegetables13.
ORAL AND PHARYNGEALCANCER
Thirteen case-control studies ofvegetable and fruit consumption andoral and pharyngeal cancer have beenreported. Ten of these studies reported an inverse association betweenoral and pharyngeal cancer and intake of one or more vegetables and/or fruits.
Consumption of fruit appearedto be especially consistent, with anapproximate doubling of risk for consumers of the least vs the most amountsof fruit. No clear pattern emerged forspecific vegetables with the possibleexception of green vegetables whichwere associated with reduced risk infive of six studies.
As with oesophageal and laryngeal cancer, the predominant risk factors for oral and pharyngeal cancerare tobacco and alcohol. In most areas of the world, smoking, chewing(including betel chewing) and drinking are estimated to account for 75per cent or more of the cancers ofthese sites. Each of the studies reviewed here found these to be themajor risk factors. Poor dentition wasalso found to be a risk factor in a fewstudies. Because, as noted before,tobacco and alcohol habits may correlate inversely with food intake andthe nutritional quality of the diet, thepotential for a confounding effect ofthese habits on the relationship between vegetable and fruit consumption and oral and pharyngeal canceris great. Nonetheless, there is ampleevidence of an independent inverse
association with vegetable and fruitconsumption.
BREAST CANCER
Thirteen case-control studies ofbreast cancer in 11 different countries have examined the potential protective effect of vegetable and fruitconsumption. Of these, nine found asignificant reduction of risk in association with at least one vegetable orfru it catego ry.
Most of the breast cancer studies focussed on fat intake as well asvegetable intake in order to shed lighton the controversial hypothesis concerning a role for a high-fat diet in theetiology of breast cancer. An interesting difference between the studies that found associations betweenvegetable and fruit consumption andbreast cancer and those that did not,is that odds ratios were adjusted forpotential confounders in the studieswhich found a relationship. Theoretically, high vegetable and fruit consumption may be associated positively with risk factors for breast cancer such as upper social class, bodyweight, and later age at first birth.Such correlations might obscure anyinverse association between vegetableand fruit consumption and breast cancerin the absence of adjustment.
Overall, the associations appearto be weaker than those seen forrespiratory and digestive cancers. Itmay be that the potential anticarcinogenic actions of vegetablesand fruits are less applicable to breastcancer than non-hormone related epithelial cancers. For instance, mostlung and upper aerodigestive cancers are strongly associated withsmoking, which provides a direct exposure to initiating agents, whereasfor breast cancer, the known risk exposures are mainly related to exposure to unopposed estrogen, whichmay have more of a promotional effect. If the effect of potential anticarcinogenic agents in vegetables andfruit is to block initiating events thenthese agents may be more effectivein preventing, say, lung than breastcancer. Nonetheless, an effect of indoles, which are found in cruciferousvegetables, on estrogen metabolismhas been demonstrated and maybe relevant to the etiology of breastcancer.
Thus, the majority of the studiesshow no more than a halving of risk
with high consumption of vegetablesand fruit. It is commonly the case thatdietary risk or protective factors arenot shown to be as strong as otherrisk factors in epidemiological studies. This may be in part due to attenuation of odds ratios due to the inability of dietary assessment methods toclassify diets of individuals with greatprecision. It may also be due in partto the narrow range of levels of exposure to dietary factors within mostpopulations.
Therefore, odds ratios consistently suggesting a doubling or halving of risk pertinent to dietary exposures may in fact be signaling a strongerrelationship. Even if the risk of cancer were only doubled by low vegetable and fruit consumption, the importance would remain because theattributable risk would be high due towidespread exposure to low vegetableand fruit consumption and becausediet is a relatively modifiable entityand thus a potential target for change.
MECHANISMS
There are many biologically plausible reasons why consumption ofvegetables and fruit might reduce thelikelihood of cancer. These includethe presence of potentially anticarcinogenic substances such ascarotenoids, ascorbate, tocopherols,selenium, dietary fibre, dithiolthiones,isothiocyanates, indoles, phenols,protease inhibitors, allium compounds,plant sterols, limonene and otherswhich are increasingly being called,collectively, phytochemicals or bioactive compounds - chemicals ofplant origin that playa crucial role inmammalian metabolism6.
When the body is first exposedto specific carcinogens, many arenot in thEilir active form; the stepsbetween exposure to the pro-carcinogen and the conversion of a normalcell into a cancer-prone cell (transformation) can be considered as follows: the pro-carcinogen is activatedto the ultimate carcinogen form byP450 enzymes. (It is worth keeping inmind that the body is not trying tomake carcinogens - it is trying tosolubilise insoluble foreign compoundsand therefore enable excretion in theurine.) However, this is complicatedby the fact that the same enzyme canoften make one compound less carcinogenic and another more carcinogenic; either of the forms of the car-
cinogen - pro-carcinogen or ultimatecarcinogen - may be converted byPhase II enzymes into a form that isrelatively inert and even more easilyexcreted.
These are typified by glutathioneS-transferase; if not excreted, the carcinogen can pass through the cellmembrane and the nuclear membrane;it can then interact with the DNAforming adducts and/or produce mutations; DNA synthesis and replication (or DNA repair) subsequentlyoccurs. Repair has varying degreesof fidelity; if the DNA is not repairedaccurately, cell replication - producingdaughter cells with copies of themutated DNA - occurs. These cellsthen synthesise an abnormal proteinor fail, altogether, to synthesise aprotein crucial to the normal functionof the cell or even crucial to controlling cell replication itself (this is almost certainly what happens when atumour suppressor gene [for example,P53] mutates or is deleted).
This sequence of stages bringsa cell a step closer to becoming acancer cell (alternatively, even withabnormal DNA, the cell may cease toreplicate and then undergo differentiation or apoptosis). DNA damageprobably has to occur several timesbefore a cell becomes completely freeof growth restraint and a fully cancerous one. Finally, the abnormal cellsobtain a growth advantage over thenormal cells and steadily increase innumbers (promotion) - often becomingmore malignant and able to spread(progression). These are steps thatthemselves involve further changesin the cellular DNA.
At almost everyone of the abovestages, known phytochemicals/bioactive compounds can alter the likelihood of carcinogenesis, occasionally in a way that enhances risk, butusually in a favourable direction. Forexample, such substances as glucosinolates and indoles, isothiocyanatesand thiocyanates, phenols andcoumarins can induce a multiplicityof solubilising and (usually) inactivating enzymes; ascorbate and phenolsblock the formation of carcinogenssuch as nitrosamines; flavonoids andcarotenoids can act as antioxidants,essentially disabling carcinogenicpotential; lipid-soluble compounds suchas carotenoids and sterols may altermembrane structure or integrity; somesulphur-containing compounds can
4
suppress DNA and protein synthesiscarotenoids suppress DNA-synthesi~and enhance differentiation.
The relation of cancer risk tcconsumption of plant foods is probably most usefully considered in anevolutionary/adaptational context. Theargument in this regard has beendeveloped in more detail elsewhere14.15.
SUPPLEMENTATION
As evidence emerged in the 1960sand 1970s of a role for vegetablesand fruits as protective agents, particularly against stomach and lungcancers, a number of investigatorsbegan to ask about potential mechanisms and agents. The animal experimental data were already suggesting the presence of a whole variety of active agents but humans havealways liked simple solutions andmagic bullets16. Researchers wonderedwhether p-carotene might be the relevant agent17• They began with retinol data on humans and animals,but drifted to p-carotene as oxygendamage to DNA and antioxidant status became possible players incarcinogenesis. This nicely framedhypothesis was enough to spark awhole industry.
The hypothesis was boldly statedand testable, and a number of wellconducted studies have now renderedthe verdict. p-carotene supplementation does not lower rates of lung cancer18.19.20and among high-risk individuals appears to increase incidenceand mortality18.19.p-carotene also doesnot reduce recurrence of adenomatous
pOlyps21 and may increase risk oflarger polyps22. ~-carotene does notreduce the risk of recurrence of skincancer23.
In striking contrast, as the constituents of tobacco smoke were enumerated, it became increasingly clearthat making a noncarcinogenic cigarette was not an achievable goal andthat behavioural, economic and legislative programmes to eliminate smoking were more appropriate. It maysuggest that behavioural approaches(paralleling the successful tobaccoprogrammes) to increasing plant foodintake may be more fruitful. Economicincentives to increase human plantfood production are also worthy ofconsideration.
There are other arguments enumerated above, based on the biology
of cancer, to support an increasedintake of plant foods as a primarystrategy. The use of single agentshas proved to be ineffectual and ultimately counterproductive in the treatment of many cancers; this is largelybecause resistant clones arise readilyin the presence of potent cytotoxic/chemotherapeutic agents. In rapidlyproliferating tissues, with elevated levelsof cell death, selection for survival inthe presence of the agent will occurrapidly. In tissues where there arelarge numbers of initiated cells, it isplausible that similar selection willhappen in the presence of a singlechemo-preventive agent.
If the action of the agent is toinduce differentiation, those cells thatare incapable of differentiation maygain a proliferative advantage. Similar arguments apply to a single agentthat increases the rate of apoptoticcell death; again, those cells that areresistant to apoptosis may continueto proliferate24.
It follows that, while we shouldcontinue experimental studies to understand the role of specific agentsin cancer prevention, we should notbe surprised if the effect of the agentdiffers at different stages ofcarcinogenesis or if it differs whenused alone vs in combination withother compounds.
The safest public health strategy seems to be to advocate increasedintake of intact plant foods with themultiplicity of agents that they contain. It is less likely that any clone ofmalignant cells can survive the polypharmacy of plant foods15,24.
IMPLICATIONS FOR INDIVIDUALSAND COMMUNITIES
The implication for individualsand societies around the world areclear - produce and eat more vegetables and fruit - not more pills.
At present, it is not clear whatquantity we should eat each day - itis clear that many people around theworld do not eat enough. Therefore,setting a population target of perhaps 400 g may be a useful interimgoal. Certainly, at this level, there arefew dangers for any part of the population. In India, maintenance of thetraditional plant-based diet, at a levelof energy intake sufficient to ensurethe absence of childhood malnutrition, is, along with the avoidance of
tobacco use, likely to make a profound contribution to continuing thelow rates of cancer that are a featureof Indian health.
Excerpts from the Gopa/an Oration delivered
at the XXIXth Annual Meeting of the Nutrition Societyof India atthe National Institute of Nutrition, Hyderabad,on November 21, 1996.
The full text of the oration will appear in the
proceedings of the Nutrition Society of India.
References
1. Kohman, E.F.: The chemical components ofonion vapours responsible for wound-healing qualities.Science, 106:625-27, 1947.
2. Fenwick, G.R., Heaney, R.K. and Muller, W.J.:Glucosinolates and their breakdown products infoods and food plants. CRC Crit Rev Food Sci Nutr,18: 123-201, 1983.
3. Darby, W.J., Ghalioungui, P. and Grivetti, L.:Food: The gift of Osiris. Vo12, Academic Press Inc,New York, 1977.
4. Doll, R. and Peto, R.: The causes of cancer.Oxford University Press, New York, 1981.
5. Steinmetz, K.A. and Potter, J.D.: Vegetables,fruit and cancer. I. Epidemiology. Cancer Causesand Control, 2:325-57, 1991.
6. Steinmetz, K.A. and Potter, J.D.: Vegetables,fruit and cancer. II. Mechanisms. Cancer Causes
and Control, 2:427-42, 1991.
7. Steinmetz, K. and Potter J.D.: Vegetables, fruitand cancer prevention: A review. JADA, 96:10271039, 1996.
8. Stocks, P. and Karn, M.N.: A cooperative studyof the habits, home life, dietary and family historiesof 450 cancer patients and of an equal number ofcontrol patients. Ann Eugenics, 5:30-280, 1933.
9. Orr, I.M.: Oral cancer in betel nut chewers inTravancore. Its aetiology, pathology and treatment.Lancet, 2:575-580, 1933.
10. Correa, P., Fontham, E., Pickle, L.W., Chen, V.,Lin, V. and Haenszel, W.: Dietary determinants ofgastric cancer in Louisiana inhabitants. J Natl Cancer Inst, 75:645-54, 1985.
11. Notani, P.N. and Jayant, K.: Role of diet inupper aerodigestive tract cancers. Nutr Cancer,10:103-13,1987.
12. Cook-Mozaffari, P.J., Azordegan, F., Day, N.E.,Ressicaud,A., Sabai,C.and Aramesh,B.: Oesophagealcancer studies in the Caspian littoral of Iran: resultsof a case-control study. Br J Cancer, 39:293-309,1979.
13. Koo, L.C.: Dietary habits and lung cancer riskamong Chinese females in Hong Kong who neversmoked. Nutr Cancer, 11:155-172, 1988.
14. Potter, J.D. and Graves, K.L.: Diet and cancer:Evidence and mechanisms - an adaptation argument. Ed: Rowland, I.R. In: Nutrition, Toxicity andCancer. CRC Press, Boca Raton, 379-412, 1991.
15. Potter, J.D.: The epidemiology of diet and cancer. Evidence of human maladaptation. Eds: Moon,T.E. and Micozzi, M.S. In: Nutrition and Cancer
Prevention. Investigating the Role of Macronutrients. Dekker, New York, 55-84, 1992.
16. Wattenberg, L.W.: Inhibition of chemical
5
carcinogenesis.J Natl Cancer Inst, 60: 11-18, 1978.
17. Peto, R., Doll, R., Buckley, J. and Sporn, M.:Can dietary ~-carotene materially reduce humancancer rates? Nature, 290:201-208,1981.
18. Alpha-tocopherol, ~-carotene Cancer Prevention Study Group. The effect of vitamin E and ~carotene on the incidence of lung cancer and othercancers in male smokers. N Engl J Med, 330: 10291035,1994.
19. Omenn, G., Goodman, G.,Thornquist, M.,Balmes,J., Cullen, M., Glass, A., Keogh, J., Meyskens, F.,Valanis, B., Williams, J., Barnhart, S. and Hammar,S.: Effects of a combination of ~-carotene andvitamin A on lung canc'er and cardiovascular disease. N Engl J Med, 334: 1150-1155, 1996.
20. Hennekens, C.H., Buring, J., Manson, J., Stampfer,M., Rosner, B., Cook, N., Belanger, C., LaMoke, F.,Gaziano, JM., Ridker, P., Willett, W. and Peto, R.:Lack of effect of long-term supplementation with~carotene on the incidence of malignant neoplasmsand cardiovascular disease.N Engl J Med, 334: 11451149,1996.
21. Greenberg, E.R., Baron, J., Tosteson, T., Freeman, D., Beck, G., Bond, J., Colacchio, T., Coller,J., Frankl, H., Hail, R., Mandel, J., Nierenberg, D.,Rothstein, R., Snover, D., Stevens, M., Summers,R. and van Stolk, R.: A clinical trial of antioxidantvitamins to prevent colorectal adenoma. N Engl JMed, 331:141-147,1994.
22. MacLennan, R., Macrae, F., Bain, C., Battistutta,D., Chapuis, P., Gratten, H.,eta/.: Randomised trialof intake of fat, fibre and ~-carotene to preventcolorectal adenomas. J Natl Cancer Inst, 87: 17601766, 1995.
23. Greenberg, E.R., Baron, J., Stukel~T., Stevens,. M., Mandel, J., Spencer S., Elias, P., Lowe, N.,Nierenberg, D., Bayrd, G., Vance, J.C., Freeman,D., Clendenning, W. and Kwan, T.: A dinical trial of~-carotene to prevent basal-cell and squamouscell cancers of the skin.N EnglJ Med, 323:789-795,1990.
24. Potter, J.D.: Chemoprevention: pl<larmacologyor biology. Oncology, 10:1487-1488, 1996.
NUTRITIONNEWS
• International Symposium by RankPrize Funds: This Symposium on:'Feeding a World Population of morethan Eight Billion People: A Challenge to Science' was held in the UKfrom December 6-9, 1996. It was attended by 104 delegates. The proceedings are expected to be published shortly.
• The 23rd Kamla Puri SabharwalMemorial Lecture was held on December 18, 1996, at the Lady IrwinCollege. Dr Prema Ramachandranspoke on: 'Current Concerns inMaternal Nutrition'. Dr C. Gopalanpresided.