Tobacco Documents Online

N Inside: Modified starch in baby foods: the selling of fear? PAGE 3 A letter to Virginia about cancer risks in the environment PAGE 8 The role of dietary i calcium in regulating blood pressure PAGE 11 Tofutti: new ice cream substitute PAGE 13 When Is a' Potential' Risk a 'Significant' Risk? What would happen if our regulatory agencies applied their views on the risks of chemicals to more personal situations, such as the risk of becoming pregnant? "There is sufficient evidence to con- clude that XXX is a potential risk to humans:' YOU CAN REPLACE THE Xs INITHE ABOVE SENTENCE with almost anything that comes to mind; this sentence-in almost exactly this wording-has been applied to an end- less list of pesticides, pollutants, food ingre- dients, forms of electromagnetic radiation, and other components of our overall envi- ronment. It has appeare& in newspapers, magazines, scientific journals, and (espe- eially) government documents. What does it: mean? What the sentence usually means is that some form of tissue reaction developed in CONTINUED ON PAGE 2 VOLUME 7, NUMBER 1 JAN/FEB 1986 PRICE: $2.00 NEWS & VIEWS PUBLICATION'OFTHE AMERICAN COUNCILON SCIENCE AND HEALTH • 1995 BROADWAY • NEW YORK, NY 10023 (212) 362-7044 Cancer Survival Rates: What the Media Haven't Told You By Alan C. Fisher and Wendy Worth FIVE YEAR CANCER SURVIVAL RATES* ONE OF THE MOST CONSISTENT MESSAGES 100 that the American media have given to the public over the last two decades is that we are winning the war against cancer. We have 90 been told that~ a combination of new tech- r nologies for early detection of the disease 80 _ 1950-54 L and improved treatments for the disease , once it is diagnosed, are resulting in increased survival'rates for cancer victims. 70- However, what is being claimed by a num- ber of respected critics is that this rosy pic- 60ture of increases in cancer survival does not f- hold up for the most serious and widespread W forms of cancer - those of the lung, colon« V so rectum, prostate, and breast. These critics ,°y~ include: John C. Bailar, a physician and bio- d 40 statistician at the Harvard University School of Public Health; Haydn Bush, the director of a regional cancer center in Canada; John 30 Cairns„a cancer analyst at the Harvard Uni- versity School of Public Health;, David Zo Eddy, a physician and mathematician at Duke University; Alvan R. Feinstein, a phy- sician at the Yale University School of Medi- 1o cine; and Richard Peto, a British epidemiol- ogist. They say that survival rates for these cancers have not really changed over the last two decades despite statistics and charts to the contrary that havebeen released by gov- ernment and private cancer organizations CONTINUED ON PAGE 6 LUNG COLON- RECTUM By Sidney Shindell 1950-54 to 1977-81 PROSTATE •Adjusted for normal life expectancy. Source: Biometry Branch, National Cancer Institute Table 1 1977-61 BREAST

When Is A Risk? CONTfNUED FROM PAGE t some species of animal in numbers exceed- ing its spontaneous occurrence when theani- maU had received as much! of the item as it can tolerate without being killed'outright by it. Often there are concomi'tanG studies showing, that no other, test species behaves similarly and that lesser amounts have no discernable effects. I have no problem if such a study stimu- lates further study in animals, biologicall models, and people, but I do get distresse& when the sentence in question becomes the sole basis for a statement which starts: "Having established the proposition that XXX presents a significant risk to human beings, we urge that the following, regulatory actions be taken . . ." Potential Risk vs. Significant Risk Somehow, without any additional evi- dence that would show that XXX is a real hazard, it has been magically transformed from a"potentiall' risk to a "significant" one. I think that we all recognize that this does not follow the rules of logic. Logically, there shoul& be a difference between "potential" andl "significant," especially when the latter word is being use&to justify some expensive andinconvenient regulatory action. Can we distinguish clearly between the two types of risks? I think we can„if we examine another biological process the way some of our confreres look at chemicals in our environment. The process that I find makes the most sense to most people is the process of pregnancy. Every female in this world, unless she has been the victim of some rare disease or has had some specific surgical! intervention, has the potential'risk of pregnancy every month of her life from the date menses start to the date of her menopause. This risk is sus- pended during actual, pregnancy and for a short period following delivery, but for the non-pregnant female, I repeat there is a potential risk of pregnancy every month during,the so-called childbearing years. Is this a significant risk? Apparentiy so, since six million pregnancies occur each year in the United States. This amounts to about 110 pregnancies per 1,000 women aged 15- 44, per year. We know, moreover, that the risk is only present if there is male contact, and only if that contact is in a very specific form under very specific circumstances. I should qualify this statement since there is a substantial body of expertopinion that on at least one occasion, conception occurred without male interventionl Basically, how- ever, I think we can all accept that the risk of pregnancy is a dose-related phenomenon. What I mean by dose-related is that the risk to a female of becoming pregnant increases whenever, there is sufficient con- tact with a male. Our society presents many situations where social contact is encour- aged, and one could certainly postulate a greater risk in these settings versus that of a 2 group of females cloistered in a convent, simply because of the opportunity for greater male-female contact. Of course, in some regulatory situations, the idea that a:potential risk might be dose- related is not accepted. In these same set- tings, regulators are not permitted to con- sider benefits that might offset risks. Food additives suspected of causing cancer are an example of a potential risk that is handled this way. Exposure to them is not merely neg- ulated, it is eliminated, and benefits are ignored. Applying this philosophy to the risk of pregnancy would mean totally elimi- nating male-female contact and ignoring the benefits thereof. Fortunately, in most regulatory settings, exposure to potential risks is merely reduced, not eliminated. Howevery even in these settings, regulators become greatly concerned if any evidence is presented that suggests that our exposure to the source of risk is increasing. Is the risk of pregnancy increasing? One could postulate a chain of events in1whichas' the result1of increasing male-female contact in business and professional settings, a more intimate relationship might develop between a specific male and a specific female, which might lead to circumstances in which the likelihood of sexual contact might increase, with the attendant likelihood of initiation of a pregnancy. Is Regulation the Answer? In spite of the fact that we could all postu- late or accept sucK a series of events and might'even construct a mathematical model to express the likelihood of a pregnancy, given the degree of contact in various set- tings, the question I have is: Would anyone today seriously feel that the intermingling of capable males andfertile females in the busi- ness setting be the subject of regulation? In other words, even though a risk of preg- nancy is theoretically increased through business contact, is the risk of sufficientsig- nificance as to be a matter of concern? I think we will all concede that at' least in this setting a sub-critical dose is generally present. Someone, I'm sure, is going to suggest that activities leading, to increasing the risk of pregnancy are on the whole voluntary, while exposures to environmental! risks, albeit theoretical at best and minimal inl amounl are involuntary, and this justifies our concern. Let me suggest: that the differ- ence between the two types of risk is not as great as it may seem. Few people consider the risks of'l environmental exposures at thee time that they assume them. Similarly, the risk of pregnancy is often far from upper, most in the mind of the individual who assumes it, at the time that she assumes it. The focus, in most instances, is on a more primary benefit expected from the male- female encounter. Risks vs. Benefits Exposure to many substances in the envi- ronment is also assumed secondary to a more primary benefit derived from their use. The risks of most exposures, however, are by no means of the magnitude of the risk of pregnancy that we are willing to accept by permitting, nay encouraging, more male- female contact in the business and profes- sional world involving meetings where both are present. We attend' these meetings knowing that the risk of pregnancy to the female attend- ees, while always a potential threat, does not occur simply because of their attendance. How many studies dome need to show that substances like saccharin or formaldehyde in the concentrations normally encountered are not accompanied' by a cancer risk to humans simply because they have been used in minimal amounts in specific situations? We have 50 years of history in the case of one of these substances and 75 years of history in the case of the other. No one has shown a human risk, and many have looked! A "potential"'risk based on animal data does not become a "signifi- cant" human risk just because someone says so. We could eliminate alli potential risk of substance X exposure, no matter how small, by eliminating substance X. We could also eliminate all potential risk of pregnancy by eliminating all male-female contact_ We have already concluded that we could pay a severe penalty if we did one of these things. What about, the other? I have much concern about even thinking in terms of risk elimination. I have yet to find any activity or any substance that doesn't constitute a hazard under certain conditions. We must, if we are going to sur- vive in this world, recognize the degree of risk in everything we do and' in everything we use, and deal with these risks accord- ingly. When all we have is a potential risk of a health effect from a substance which is not expressed in the quantities in which we use it, we must finally agree we are not subject- ing ourselves toa hazard: Animal studies are useful to focus our attention. They do. We look to see whether there is a human hazard. When we find one, we control it. When~we don't, we keep look- ing. The questionl when do we stop looking and go about our business?' fla Sidney Shindell, M.D., LL.B., is Professor and Chairman, Department of Preventive Medicine, the Medical College of Wiscon- sin. He has been a member of ACSH's Board ofScientificAdvisorssince 1978, and serves on the National Advisory Committee on Occupational Safety & Health. He is a Regent of the American College of Preven- tive Medicine and a consultant in occupa- tional and environmental epidemiology.

Modified Starch and Baby Foods The Selling of Fear? ' By Linda M. Behre MANY PARENTS TODAY ARE CONCERNED with their baby's nutrition - and with good reason. Nutrition does indeed play a vital role in the proper growth and development of infants an&childten. But parents need to be aware that baby foods, like most other foods available today, are competitive prod- ucts. And nutrition is increasingly being used as a marketing tool to give foods a com- petitive edge. A perfect example of this strategy is cur- rently being used in a new line of baby foods from Beech-Nut. The recent focus is on modified starch. The implications center primarily upon the safety and digestion of this widely used food ingredient. Chances are, many parents are not highly knowledgeable of the role of modified starch in food systems. Those who are unaware may consider it just another "unnatural" food additive that babies cer- tainly do not need, especially one's own. This is precisely the attitude that Beech-Nut is attempting to create through current advertising. Parents are being led to believe these new baby foods are better for baby, because they do not contain the allegedly hazardous mod- ified starch. Is there some scientific merit here, or is this just another case of the selling of fear (see ACSH NEWS & VIEWS May/ June '82), leading mothers unnecessarily to eliminate modified starch products from their babies' diets? Look Mom (and Dad) No Modified Starch! Why all the recent publicity? Beech-Nut has decided to remove "chemically" modi- fied'starches from some of its products. In addition to extensive advertising on T. V. and in parents' magazines, a"pediatria mes- sage"' in the form of a pamphlet has been released'for distribution to assist in the pro- motion of a new line of baby foods. The pamphlet, which offers many facts about modified starch, explains why babies will benefit from the removal of modified starch. - We feel this information deserves clarifi- cation in an effort to help parents make an informed decision regarding the purchase of baby foods. What Are Modified Starches? Native or unmodified starch is a food. However, in many applications in the food industry - particularly those involving the use of cooked'starch - it faces serious limi- tations. These include decreased stability, shortened shelf life and unpalatable texture - changes which result in reduced: Parents need to be aware that baby foods,Jikemost other foods availableloday, are competitive prod- ucts. and digestibility. Modified starches were developed primarily to overcome these shortcomings. Starches prepared from a number of sources (see chart) are considered "modi- fied"' when they have been chemicallyy treated to introduce substituent groups onto the starch and/or by forming crosslinks between neighboring starch molecules. For this reason Beech-Nut refers to modified starch as "a chemical alteration." LL Recent efforts used in pro- moting a new line of baby foods from Beech-Nut focus on modi- fied starch-the safety and digestion of this widely used food ingredient. ~9 While most of the food we consume has undergone chemical alterations, it remains perfectly safe and nutritious. Statements such as this often serve to reinforce the wide- spread negative attitude individuals hold regarding additives or chemicals. Through modifying treatments, signifi- cant contributions have been made to the food industry. Altering the physical and chemical characteristics of native starches enables them to meet our high standards in terms of food quality, processing, packag- ing, and distribution while essentially retain- ing the appearance, texture and nutritive value of the native starch. Modified' starch acts as a stabilizer, an essential component ofl baby foods, required to suspend the finely divided food particles. Prior to the availability and use of modified starch, there was concern because baby foods were too cohesive and lumpyy which led'to a gagging tendency. Early baby foods were thickened with potatoes, flour and/or carrots. New products are replacing modified starches with carob gum as the stabilizing agent. The use of nutritive stabilizing sys- tems containing modified starch is preferred over the use of alternate non-nutritive gums, thickeners or man-made systems. By modifying starch, the food manufac- turer can use significantly lower levels of starch, than might be possible with native starches. For, example, the amount of unmodified tapioca starch required to attain approximately the same thickness as modi- fied tapioca starch in strained beef and vege- tables results in a 56 percent reduction in starch level. The advantages of reduced starch for infant nutrition are a decrease in caloric content with, an accompanying increase in nutrient density. This reduction also plays an important role in infant diges- tion, as will be discussed. How Safe Are Modified Starches? Beech-Nut's promotional material for its new baby food line implies that parents who feed their children food containing modi- fied starch are gambling on safety. This is CONTINUED ON PAGE 4 3

Modified Starch CONTINUEDFROMPAGE3 Use of Modified Starch in Baby Foods Major Modified Starches Purpose of Modification Source of Starch Food Class • acetylated distarch • provide desirable • corn (waxy maize) • strained and phosphate texture junior f'oods • cassava (tapioca) • distarch phosphate • thicken, improve • potato • dinners • acetylated distarch palatability • wheat • high meat dinners adipate • extend shelf life • desserts and fruits • stabilize foods • vegetables during processing, distribution and storage • reduce levels of starch • reduce nutrient losses • increase microbial safety simply not true. The use of modified food starch in baby foods has been subject to rigid scrutiny. All modified food starches were extensivelyy examined and reviewed by the Food and Drug Administration (FDA) prior to being included in the GRAS (Generally Recog- nized as Safe) list of food additives. Limita- tions in treatment and residues as well as lev- els of substitution have been diligently established and monitored'bythe FDA. In 1970 a subcommittee to evaluate Safety and Suitability of MSG and Other Substances in Baby Foods reported, "Avail~ able information indicates that there is no toxicity, even at levels of intake substantially above use, with any of the modifications currently approved for food use in the United States." In addition, "there is no basis to conclude that the currently used modified food starches are less well metabo- lized than the native starches from which they are prepared; however, there is lack of clinical'evidence of this point." The findings of the FDA subcommittee are noted by Beech-Nut to have been released 15 years ago. Has recent evidence disproven the safety of modified starch? In 1978 an Expert Committee for Evalua- tion of Modified Starches in Infant Foods convene& by the American Academy of Pediatrics, under contract to FDA, released similar findings. The report concluded, "There is no evidence that modified food starches, at current use levels, are associated with any clinical problems. However, because infants younger than three monthss may have limited ability to digest modified and unmodified starches, it is recommended that total dietary starch use be limited in minimal amounts as practicably possible in this age group:" Since 1978, there have been several stud- ies which agree with the recommendations of the Academy. Following its conclusions, it seems clearly advantageous to use modi. fied food starches in place of unmodified starches since the same effect can be accom• plished using less starch. No modified starch may be produced using a treatment other than what is now permitted until its safety has been demon« strated. According to the , NAS-NRC (National Academy of Sciences-National Research Council), "there appears to be no toxicologic basis for excluding the use of modified food starches from the diet of infants ... provided the degree of modifica- tion, as defined by regulation, is enforced and the amounts used to achieve the benefit do not exceed good manufacturing prac- tice." In addition to the FDA and the American Academy of Pediatrics, the Scientific Com- mittee for Food of the European Economic Community has reviewed the research extensively. It concludes in "Reports of the Scientific Committee for Food" that modi- fied starches (including those used in baby foods, distarch phosphate, acetylated di- starch phosphate) can be regarded as accept- able. The Committee also considered it "unnecessary to establish ADI's (acceptable daily intakes) provided technological usage remained at present-day levels .." which indeed is necessary since higher amounts would'result in anunpalatable product. The Joint Expert Committee on Food'Additi'ves of the FAO/WHO has also reviewed the latest findings and continues to consider modified starches suitable for use in foods without specifying an ADI. This has reaffirmed the position taken by regula- tory groups in the ULS., Canada, and the U.K. that, "Modified starches impart valu- able functional properties to foods and have nutritive value. They consist essentially of starch~wifflow to very low levels of substi- tuent groups. Extensive safety studies reviewed by independent regulatory bodies provide adequate assurance of their safety." Can Infants Digest Modified Starches? A major point emphasized in recent advertising is the question of digestibility of modified starch. Beech-Nut's new line con- sists of baby foods physicians and nutrition- ists can "trust." Are modified starch,con- taining baby foods not to be trusted? Interest in starch digestibility in early infancy arises from two main concerns: first, to what extent starches can be consid- ered as a caloric source and second, what is the potentialirisk of undigestedstarches pro- ducing diarrheal symptoms and malabsorp- tion? It is true that young infants and children have some difficulty digesting starch (modi- fied and unmodified); however, reports of modified starch intolerance are rare. Beech- Nut discusses the modification technique, crosslinking, using distarch phosphate and acetylated distarch phosphate. It is stated that' modification decreases the digestibility of starch. There is no mention about degree of modification. The point to be made is that any starch, modified or unmodified, in large quantities will cause digestive problems, such as diar- rhea and malabsorption of nutrients. Many safe chemicals in excess coul& pose health problems. However, the amounts of acetyl and phosphate groups used'in crosslinking are strictly regulated4 as are all chemicals in our food supply. Certainly if the evidence revealed a clear digestibility problem as a resuloof modified starch use in baby foods, it would be removed from all baby foods. Although the FDA report released in 1978 found no evidence for revising pre- vious evaluations, it did recommend addi-

tional studies on the digestibilityof complex carbohydrates (starches) by extremely young infants. Many of these studies have been completed. Digestibility studies pub- lished since 1983 have verified that the caloric value of modified starches used in baby foods is similar to their non-modified counterparts. Starch requires a combination of enzymes for its complete digestion to simple sugars. The key enzyme in the hydrolysis, or break- down„of starch is pancreatic amylase. Most questions being raised concern the ability of the young infant to digest large quantities of starch due to low levels of pancreatic amyl- ase. An adequate group of enzymes for all simple nutrients except starches is appar- ently available before birth. Pancreatic amylase, therefore, is age-dependent and both clinical and experimental investiga- tions in young developing infants are extremely limited. Up to the age of 6 months, it has been shown that pancreatic amylase, although low in infants, may still be sufficient to digest dietary starch. Most newborns and young infants do tolerate moderate amounts of starch, but the tolerance of young infants to very large quantities of starch is limited. One investigation, concluded that rarely does a deficiency of pancreatic amylase lead to clinical symptoms in infants. It found that normal infants, 1 to 3 months of age, will tolerate without difficulty 10-15 pereenG of total dietary energy as starch. Few infants will be fed diets providing this level of starch during the early months of life. Dr. L. J. Filer„Jr. of the Department of Pediatrics, University of Iowa, and the Chairman of the FDA subcommittee which reviewed the safety of modified starch in baby foods, stated, "the quantity of dietary starch ingested by infants, children and adults in the U.S. is well within the func- tional capacity of the gut." A study in 1983 concluded that young infants cam utilize cereal, although absorp- tion is not'always complete. Similarly, a 1982 study of food intake and' growth pro- vided convincing evidence of the ability of infants to digest starch (unmodified) ade- quately after 6 weeks of age. This study sug- gested, however, that some infants less than 6 weeks of age had limited ability to digest starch. In contrast, some research suggests that pancneatic amylaseactivity is not detectable in infants until 4 to 6 months of'~ age. The conflicting results of these studies have led to controversy regarding the age at which starch should be introduced into the diet of inf'ants. Some of the comparative rates of diges- tion for modified versus unmodified starches were not established in 1978. Since then, studies of rates, outside the body (in laboratory hydrolysis studies) as well as inside the body„have been reported. Twenty years ago, starch digestion was studied through laboratory techniques out- side the body. However, carbohydrate diges- A study in 1983 concluded that young infants can utilize cereal, although absorption is not always complete. tion and absorption are now known to be part of a much larger metabolic process, which requires studies involving the whole organism. Animal studies are frequently used for this purpose. Neither method, how- ever„ can replace the complex 'human ~ diges- tive system. While modification changes the physical properties of starch, it has a minimal effect on the manner in, which the body handles these starches. Studies on distarch phos- phates (substituent group added to starch during modification) indicated no signifi- cant differences in digestibility between modified and unmodified starch. The level of substitution is too lbw for the starch to be significantly different from unmodified starch in the digestive process. In experimental animals, when a large part of the total diet, 50 percent or, more in many cases, is composed of modified starches, there may be evidence of diarrhea and cecal enlargement, particularly in the very young animal. It is generally accepted by toxicologists that cecal enlargement is the result of overloading the diet with slowly digested materiah These reactions, however, do not appear at low dietary levels and rarely appear below the 25 percent level. What Are Current Intakes of Modified Starches by Infants? Modified star& contributes 10-40 per, cent of the calories provided by some infant foods. However, the contribution of these foods to the diet is generally limitedl and market research suggests that modified starches rarely account for more than 100 percenuof the caloric intake of infants. In a series of independent surveys, the intake of modified starch by infants in the first year of life was found to range from 2-6 percentof total energy intake. In one survey of 430, 1-14 month old infants, the average daily, intake of modified starch was found to be about 2=3 percent of the total'caloric intake. The maximum amount of modified' starch~consumed was about 16 percent and only four of the 430 infants obtained more than~l0percent of total calories from modi-_ fied starch~ The American Academy of Pediatrics investigation found that modified food starch contributed less than eight percent of the calories. And according to a NAS survey of GRAS substances, modified food starch provides from 24 percent of the total daily caloric intake. This low level of contribution to total die- tary intake emphasizes the fact that modi- fied starch mainly fulfills a technical rather than nutritional role in the diet'. It is gener- ally agreed that these levels should not give rise to any health problems. Long term studies concerning the utiliza- tion of starch by young infants have not been reported. It must be stressed that there is a need for additional data. This lack of clinical evidence or long term studies in normallinfants is not a valid cause to assume the worst,. Because infants younger than three months of age have a limited ability to hydrolyze starches, it is rec- ommended by the American Academy of Pediatrics that starch use, modified or unmodified, be limited to as minimal amounts as practical in this age group. Recommendations So the question exists - should parents avoid the use of modified' starch in baby CONTINUED ON PAGE 9 5

Cancer Survival Rates CON77NUEDFROMPAGEI A physician/biostatistician and a cancer analyst, both at Harvard Uni.versitySchool of Public Health; the director of a regional cancer center in Canada; a physician/mathematician at Duke University; a physician at Yale University School of Medicine; a British epidemiologist-and the American Council on Science and Health-believe that the rosy picture of increases in cancer sur- vival does not hold up for cancers of the lung, colon-rectum, prostate or breast: and publicized by the media. The American Council on Science and Health believes that the points these critics are making are valid. Table I (on page 1) shows the five year survival rates for lung, colon-rectum, pros- tate, and' breast cancers for a recent period and an earlier period. The crux of the dis- agreement centers on how to interpret cur- rent cancer survival statistics such as these. Though these kinds of statistics appear objective, their meaning is not. Critics of the current optimism on cancer statistics have given six reasons, which we will discuss in detail below, as to why they believe the cur- rent statistics are invalid and'should not be compared to past statistics. All of these rea- sons argue that current statistics contain bias; that is, they do not reflect the true state of affairs. They misrepresent cancer sur- vival so as to make it look longer than it is in actuality. Lead-Time Bias Today, because there have been great advances in technologies for cancer detec- tion, growing numbers of individuals have their disease diagnosed before they have any symptoms. Cancers in this stage are called preclinical. In the past, preclinical diagnosis would not have been possible. This early diagnosis is usually made through screening programs. A screening program is a simple medical test which is given to large numbers of individuals who generally do not have symptoms of the disease for which they are being screened. Probably the best' known screening procedure is the Pap test for cervi- cal cancen Once screened, individuals can be classified as to how likely they are to have the disease. If they are classified as likely to have the disease, further diagnostic tests can confirm its presence. Cancer survival rates are generally reported to the public in terms of the per- centage of people who live for at least five years after their disease is diagnosed. (Although other measures of survival are often available, they are rarely reported by the media.) These rates are not cure rates and do not reflect the quality of survival. Whatearlydiagnosis does is to inflate these five year survival rates through the process of lead time. Lead time is the extra time added on to a patient's survival which is due solely to early detection of the disease and not to a later time of'~ death. By turning the clock back for the time of diagnosis, the sur- vival of the individual will appear to be longer. Therefore, critics claim iU is not fair to compare survival rates of today, which contain cases detected by screening, to those of the past which do not contain suchicases. Screening favorably biases cancer statistics. The crucial issue is not whether there is lead-time bias because even the optimists who paint a rosy picture agree that it exists. The real disagreement centers around how much of the improvement in survival rates is due to just setting the survival clock earlier and how much is due to the head start that early detection gives physicians in treating the disease and also using improved meth- odfi of treatment which actually result in a later time of deaths The first result would be due to what is called a"st'atistical, artlfact,"(something created by the way the statistics are computed and not a reflection of reality) and the second would be a real difference. Critics argue that lead-time bias accounts for most or even all of the increase in sur- vival rates. Their position is being supported by an increasing number of scientific studies designed to answer this very question. So far it has been shown that lead-time bias does L L The reason why people with slow growing cancers are more likely to have their disease detected in a screening program is that these cancers, because they do grow slowly, exist in a person's bod»for a longerper- iod of time before they can be detected by clinical means- gro wths and other symp- toms. 55 account for most of the improvement in lung and breast cancer. There is also evi- dence that this bias is inflating the improve- ments imsurvival statistics for bone cancer (osteosarcoma). Length Bias Sampling A second reason why cancer survival sta- tistics are not valid is "length bias sam- pling." This kind of bias is a type of "prog- nostic selection bias: " Prognostic selection in this case refers to the fact that individuals who have their cancers detected by screening do not represent the "average" cancer patient. They have a better outlook for sur- vival because there is something different about them. In length bias sampling the claim is that the individuals who have been diagnosed'as having cancer by screening programs are more likely to have slbw growing cancers. Since people with slow growing cancers are more likely to survive longer, this type of individual artificially inflates survival sta- tistics. It has been demonstrated that slow growing tumors are often more likely to be detected in screening procedures. The reason why people with slow growing cancers are more likely to have their disease detected in a screening program is that these cancers, because they do grow slowly, exist in a person's body for a longer period of time before they can be detected by clinical means - growths and other symptoms. The fact that the time frame during which they are "silent" is longer gives the screening proce- dure a better chance to detect them before any symptoms develop. The point being made is that these very same people, even if their cancer was to go undetected, would survive longer than most cancer patients anyway because cancers that are slow growing in a preclinical stage are usually slow growing in a clinical stage. Therefore, these people bias the survival data by making it appear that early detection causes them to have survived longer. It is thus argued that it is the nature of their dis- ease - slow growing - that~ causes them to have longer survival and not early detection. This is a separate issue from lead-time bias because individuals with slow growing cancers would' be expected to live longer even if' their cancers were detected after symptoms appeared. So length bias is saying that comparing the survival'lof people whose malignancies are detected by screening with those who are detected by clinical means is like comparing two different diseases; slow growing ones with a more positive prognosis versus fast growing ones with a more nega- tive prognosis. This is what comparing present survival statistics with those of the past is really doing. Overdiagnosis A related type of prognostic selection bias is "overdiagnosis." The claim here is that screening programs can detect cancers that would never have reached the clinical stage - produced signs and symptoms - although technically they are cancers. It is possible to have such cancers and never know about' thems or never be harmed, because the tumors either do not grow or because they regress (get smaller) on their own. Since screening programs are the way individuals with these types of cancers are detected, similar individuals were not counted in cancer survival statistics in, the 6

past. However, they are counted today and thus favorably bias survival rates. Today's overdiagnosis of prostate cancer, for exam- ple„accounts for greatly increased five year survivalirates compared to the past. Patient Self Selection An additional type of prognostic selec- tion bias is "patient self selection" or "vol- unteer selection." People often volunteer to be screened. It is claimed that those who vof unteer for screening programs and those who do not get screene&are different. Peo- ple who. volunteer are more likely to be health conscious and may even be more s inely attuned'to more subtle signs of cancer. One example of such a volunteer effect was in a well known breast cancer study funded by the National Cancer Institute and con- ducted' in New York on women covered byy ihe Health Insurance Plan of Greater New York. It was found that in the "test" group (those given encouragement to be screened)) the women who volunteered for screening were more interested in their health and bet- ter educated. Such people are likely to be more compliant with physicians' orders,, during the treatment of their disease, if they are diagnosed as having cancer. Thus, these individuals would have a better outlook for survival regardless of a screening procedure. Because it is also claimed that people who are healthi conscious usually have a high, degree of education and a high income, one could therefore argue that they would have a better prognosis. They would be able to obtain higher quality physician, care and health care services, not' because their dis- ease was detected earlier, but because of their resources. Stage Migration A fifth argument relates to the different way in which, cancer patients are classifie& today as compared to the past. It is called' "stage migration." Each type of cancer can be classified into stageswith stage li being the least advanced and stage 4 the most advanced. The critics of the optimistic picture of' cancer survivali d'o not saythatithe classification scheme itself is causing the difference (although subjectivity may be involved in classification to some extent) but that new technologies to detect the spread'of the disease are more sensitive than the technologies used in the past. The additional informtition these technologies give to physicians often causes them to put newly diagnosed cancer cases into different staoes than they would have a generation aeo: Whereas lead-time bias is the result of tl~e initial early diagnosis of cancer, stage mireration bias is the result of early detec- tion of the cancer's spread (metastases). However, stage migration, unlike lead- time bias„does not influence the overall sur~ vival rate of victims of a certain type of can- cer. IU does influence the survival of victims within each stage and when statistics are reported which reflect survival rates by stage this type of bias is relevant. The survival rate of each stage is influ- enced by reclassifying individuals whose cancer is determined'to be spreading. For example, because of new technologies to detect spread, individuals who previously would have beem classified in stage 1 may now be classified into stage 2. In the pasti the spread of the cancer coul& not have been detected. This reclassification causes the survival of stage 1 to improve because the most serious cases are now being removed from this group. It also causes the survival of stage 2 to improve because the cases being added to this stage are less serious than those who would have been there in the past. Their spread is in its very early stages whereas in the past such cases would remain silent andl stayin stage 1. Thus, within eachstage a bias is being introduced. A recent article in the New England Jour- nal of Medicine(Feinstein et al., 1985) dem- onstrates this process for lung cancer. How- ever, as of yet little has been done to study this phenomenon for other forms of cancer. L LWhat the critics are claiming is that it is in the area of the more common and serious can- cers - lung, colon-rectum, prostate and breast-that the statistics are deceiving. " Increased Reporting of Non-Fatal Cases A final argument which questions the validity of current cancer survival statistics concerns the greater compliance on the part of physicians and other health professionals in reporting and registering non-fatal cancer cases. It is claimed that at the same time the recording of cancer deaths has not changed'. Thus, again it seems that the statistics have improved when it is possible that the real survivali rates have not changed, or have changed less than statistics claim. This is so only because non.f'atal cancer cases were underrepresented in the past but are less likely to be so now. Has Any Progress Been Made? The information presented above raises two questions. First, "Does this mean that' nothing has been done for cancer patients and efforts at early detection are useless?"' And second, "Why have the media not informed the American population about the controversy regarding the validity of cancer survival rates?" It is by no means true that there has beeni no progress in treating and'even curing can- cer. Great strides have been made in areas such as childhood leukemia and Hodgkin's disease, for example. But these are among the rarer cancers and what the critics are claiming is that it is in the area of the more common and serious cancers lung, colon- rectum, prostate, and breast - that the sta- tistics are deceiving. However, since infor- mation is not available as to what percentage of cancers are now detected by sereeningand no one knows the extent to which the statis- tics are being distorted by the six kinds of biases discussed above, it seems reasonable to assume, for the present, that some of the improvement in cancer survival rates is real. It may be a lot or it may be a little, we just do not know how much rightnow: We can say with certainty that some of the apparent improvement is due to biases but that doesn't necessarily mean that the cancer survival picture is bleak or that screening procedures should be avoided. The likeli- hood is high that some of the improvement is real and if this is the case screening may be saving lives. Inithe meantime what is needed are more scientific studies to determine the amounts of bias produced' by the different reasons outlined above and to determine if they are uniformly important for all forms of cancer. Equally important is the issue of why the media have not publicized this controversy.' Im general, the media receive medical news from press releases: in the present case from the press releases of government and private cancer organizations. The inherent problem here is that the organizations writing the press release choose what the media are to be told, and those organizations are unlikely to report news which is against their own inter- ests. If cancer organizations were to release information seriously questioning improve- ments in cancer survival, they would lose public confidence and government and pri- vate contributions. People do not like to - give money to a sinking ship, but they do like to contribute to a: worthwhile endeavor. Thus, these organizations have a vested interest in presenting a rosy picture even though they themselves are aware of the crit- icisms of their survival statistics. In addi- tion, cancer organizations are aware that pessimistic news would discourage many cancer patients currently in treatment. This could cause a change in the patient's atti- tudes and behavior, both of which might result in an earlier death. The distortion in reporting is carried fur- ther by the media which create their own bias by tending to present cancer survival news in a sensational way and failing to dig deeper and!question~the claims of the organ- ization writing the press releases. As con- sumers of medical news, you have the right to question~ this type of reporting and demand a more balanced view of the issues. We thus feel it is your obligation to directly question the media regarding stories slanted toward presenting only an optimistic picture of cancer, survival. y lt, Alan C. Fisher Dr. P.H., and Wendy Worth, Ph.D., are consultants in research method- ology and statistics.

Editorial , Yes, Virginia, You Are Surrounded by Carcinogens Dear Virginia: I am following up on our recent conversation wherein you expressed your deep concern about the cancer risks posed by various man-made chemicals in our environment, including Red Dye #3, dioxin, and EDB. You pointed out to me that every day you see reports on how these chemicals cause can- cer in laboratory animals and you feel that our government regulatory agencies, because they are not moving to elimi- nate these substances from our lives, appear to think "a lit- tle bit of cancer is O.K." I recall your outrage as you reminded me that some 480,000 Americans will die this year from some form of malignancy. You seemed impatient and indignant about what you see as a recent explosion of cancer causing agents in our air, water, food, and general environ- ment and what you perceive to be a parallel increase in the cancer toll in this country. Your concern about cancer is understandable. It is the second leading cause of death. But let's keep some perspec- tive here. Contrary to what some media reports have claimed, during the past' few decades of rapid industrializa- tion there has been no general increase in~ cancer rates. Indeed, only two important changes in cancer mortality can be observed: a spectacular increase in lung cancer (almost exclusively attributed to cigarette smoking) and an equally impressive decline in mortality from~ stomach cancer, most likely due to improved means of preserving food. i L Contrary to what some media reports have claimed, during the past few decades of rap id industrialization there has been no general increase in cancer rates. 5 5 But even though there is no increase, cancer remains a major public health problem, particularly when it causes suffering and death to the young and those in the prime of life. Cancer prevention must be high on our list of priorities. But obviously, we want to concentrate our preventive efforts on things that will work. There is no point in track- ing down false leads. This brings us to the question of what role do food addi- tives, pesticides and 'cont'aminants in the general environ- ment play in the causation of human cancer? Here we have to turn to mainstream science for the answer. The state-of- the-art review on this subject published recently in the Jour- nal of the National Cancer Institute confirms that while there have been a relatively small number of cancer deaths caused by high exposure to chemicals in occupational set- tings, "chemicals" like EDB and Red Dye #3 in our general environment play no known role in human cancer causa- tion. So why are you so concerned? You are probably thinking that while it may not have been proven that various man- made chemicals cause cancer, it is possible they do, since 8 many of these chemicals cause cancer in animals. Let's think that through. Are you suggesting that we eliminate exposure to every last trace of anything that causes cancer in animals? If we were to proceed in that direction, we'd have quite a job. For example, a naturally occurring constituent of nut- meg, cinnamon and black pepper causes liver cancer when administered to rodents. Tannins, which occur naturally in coffee, tea, red wines and many other foods, cause cancer in rats and mice. Some of the most popular types of mush- rooms eaten in this country contain substantial amounts of compounds in the hydrazine family of chemicals, and many of these are potent carcinogens in animal tests. Aflatoxins, a group of chemicals produced by natural molds that grow on peanuts, corn and other grains, are potent carcinogens in animals. The more we test naturally occurring chemicals in food-the more carcinogens we find. LL What is important, Virginia, is not whether or not a chemical causes cancer in lab animals, but how strong a cancer-causing agent it is, how many different types of animals respond nega- tively to it, how much we are exposed to it-and how vital a role that chemical plays in our lives. 51 What is important, Virginia, is not whether or not a chemical causes cancer in lab animals, but how strong a can. cer causing agent it is, how many different'types of animals respond negatively to it, how much we are exposed to it- and how vital a role that chemical plays in our lives. If, for example, an artificial sweetener induced cancer very quickly in animals and/or caused tumors in dogs, guinea pigs, or mice in addition to rats, and we consumed lots of it, prudence would probably dictate that we use some safer alternative or at least reduce our consumption of it. Judg- ment calls are made by scientists and regulators all the time with a full understanding that extremely low levels of toxins or carcinogens pose no known hazard. For example, the Food and Drug Administration acknowledges it is impracti- cal to eliminate all the traces of naturally occurring aflatox- ins-strong animal carcinogens-from peanuts. But pru- dence led the FDA to set strict but reasonable limits as to how much aflatoxins are allowed in our food, that is, levels that pose no cancer risk to consumers of peanuts. The point is that we should be equally reasonable about accepting exposures to innocuous trace levels of man-made chemicals known~to be animal carcinogens. Even conceding that nature is not benign, you may ask "why add to the naturally existing background of carcino- gens?" That, Virginia, is an emotional, not a rational argu- ment. If you are going to take action against chemicals that cause cancer in animals, their source should be irrelevant. Thus, you would then have to ban pepper, nutmeg and cin-

Li The more we learn about the omnipresence of nature's carcinogens, the more we realize that banning a chemical, like EDB, on the basis of `'ust in case' is like removing a grain of sand ,rom a beach - a beach thatposed no known hazard in thefirst place. 19 namon, as well, which is clearly ridiculous since these condi- ments pose no known hazard to human health. The more we learn about the omnipresence of nature's carcinogens, the more we realize that banning a chemical like EDB on the basis of "just in case" is like removing a grain of sand from a beach-a beach that posed no known hazard in the first place. If you were to demand that literally every detectable trace of all carcinogens be removed from~the environment, i L Yes, Virginia, you are surrounded by carcino- gens, natural and man made. But the level of exposure is so minimal, it is of no conse- quence. 55 there wouldn't be much of anything left. Detection tech- niques are now so sensitive we are approaching the time when we can find trace amounts of anything in everything. It has been said that in every breath of air we inhale, we take in a few molecules of the last gasp of Julius Caesar when he uttered "Et tu, Brute:' You might argue, there are no risks associated with ban- ning a chemical, so why not do it? But there are very real risks here. If we keep banning pesticides because they cause "Yes, Virginia, You Are Surrounded by Carcinogens," was published in the Los Angeles Times, October 14, 1985, while News & Views was in press. Dr. Elizabeth M. Whelan is Executive Director of ACSH. Her latest book is Toxic Terror, published in October 1985 by.7ameson Books. Modified Starch tal needs for different ages. Charts are being CONTINUED FROM PAGE 5 foods and instead'buy new products with no modified starch? No. At least not because it's "unnatural:" Manufacturers of baby foods have been using modified starch for 33 years. It has undergone extensive evaluations since 1970 by experts from the National Academy of Sciences, the Food and Drug Administra- tion, and the American Academy of Pediat- rics. While some parents prefer to feed their young children foods without anything provided which depict the nutritional improvements of Beech-Nut's new line. LL Developing an awareness of the current marketing strategies for baby foods and refusing to be swayed by a fear of perfectly safe additives would be a wise `first step' for parents. 5Y added - sugar, salt or modified starches - f or commercial baby foods do save Ume busy parents. And because of fast and closely regulated preparation techniques, commercial baby foods are sometimes even more nutritious and safer than their home- made versions. Increasing the amounts of vegetables and fruit by using less water„sugar, and modified starch is an excellent idea. So is the new con- cept developed by Beech-Nut for specially designed foods which meet the developmen- However, creating unnecessary concern over a safe ingredient is not the way to pro- mote a product. Let its merits speak for themselves. Parents should not forget that, in addi- tion to starch, vitamins, minerals, protein and fat are all important nutrients necessary for growing babies. Starches such as rice, potato and corn- cancer in animals, we are going to have less food at higher cost. The food that remains will be vulnerable to the type of insect contamination that can make us very sick. Further, given that our public health dollars andtime are limited, any effort focused on non-risks is by definition diverted from dealing with known cancer risks. LL In short, Virginia, I am not in favor of a little bit of cancer! I am, however, in favor of a little bit of common sense. 9 9 Yes, Virginia, you are surrounded by carcinogens, natural and man-made. But the level of exposure is so minimal it is of no consequence. Forget these hypothetical risks and con- centrate on the known ones. If you don't smoke, use alcohol in moderation, eat a balanced moderate diet, avoid exces- sive exposure to sunlight, obey occupational guidelines if you work in a chemical setting, and avoid promiscuous sex- ual activity, you will have done everything in your control to reduce your risks of cancer. In short, Virginia, I am not in favor of "a little bit of cancer." I am, however, in favor of a little bit of common sense. Sincerely, Elizabeth M. Whelan starch, are not bad in themselves, but they may dilute the vitamin and mineral content. In general, it is better to choose plain vegeta- bles and avoid creamed varieties. This way your baby will get a greater percentage of the Recommended Dietary Allowances (RDA) for vitamins and minerals. The same idea goes for high-sugar baby fruits which could be dilute in nutrients. Limit the use of fancy fruit "cobblers," "desserts," and "supremes" because they are usually higher in sugar with few essential nutrients. When it comes to buying baby foods, the less sugar, corn syrup, modified starch and' added flour, the more nutrients you get for ~ the calories. - Developing an awareness of the current c" marketing strategies for baby foods and ~ refusing to be swayed by a fear of perfectly 3 safe additives would be a wise "first step" ~ for parents. +~;; ~ r T ~ z Linda M. Behre, M. S., is a Research Associ- ~ ate with ACSK < 9

Profile Meet ACSH Advisor Eric W. Mood, M.P.H. Associate Clinical Professor of Public Health Yale University School of'Medicine LL Too much emphasis in public health today is being placed on the provision of health care and medical services rather than on preven- tron of illness, disability and premature death. y5 Education: 1943, M.P.H., Yale University; 1938„B.S., University of Connecticut. Research Interests: "My principal research over the past few years has been in the area of environmental epidemiology. At the present time we are midway through a study of the health effects of water fil- tered through granular activated carbon filters. I am directing this study, which involves families living in government housing at the U.S. Navy Submarine Base, Groton, Connecticut, for the U.S. Environmentai Protection Agency. "During the past five years I havee directed studies which examined the health effects of swimming in waters which may have high bacterial! densities but which are free from known sources of sewage pollution. Another series of studies looked'at the role of the quality of swimming water in cases of otitis externa, or 'swimmer's ear,' as it is more commonly called. "I have also been engaged in the devel- opment of'reeommended legislation con- cerning the design, construction, opera- tion and maintenance of swimming pools and bathing places. I have chaired for several years the Public Health Joint Committee on Swimming Pools and Bathing Places and'the American Public Health Association has published these recommended codes. "Also, I have been active in housing and health issues. In~ 1975, I chaired the committee which developed and pub- lished the APHA-CDC Recommended 10 Housing Maintenance and Occupancy Ordinance. At the present I am chairing the committee which will be publishing a new version of this recommended legisla- tion. The new document will be calle& Housing and' Health and will include other material in addition to the model' legislation." Turning Point: "I cannot identify any spe- cific turning point in my careen My pro- fessional life has been guided by the prin- ciples of two of my professors, the late Professor Charles-Edward A. Winslow and Professorlra V. Hiscock. I have been most fortunate to have had these persons as my mentors." Personal Health Code: "With the exception of smoking, my personal health code may be summed up in one word - mod- eration. Neither I nor any of my family smoke. I have always been a non- smoker." Most Important Public Health Concern that Americans Face Today: "The pre- vention of ill-health. Too much emphasis inpublic health today is being placed on the provision of health care and! medical services rather than on prevention of ill- ness, disability, and premature death. Our society, led by physicians, claims to be oriented toward primary health care but fails to acknowledge openly that pri- mary health care includes 'education concerning prevailing health problemss and the methods of preventing and con- Eric W. Mood trolling them; promotion of food supply and proper nutrition; adequate supply of safe water and basic sanitation;,maternal and child health care, including family planning; immunization against the major infectious diseases; prevention and control of locally endemic diseases; appropriate treatment of common dis- eases and injuries; and provision of essentialldrugs.' " Most Important Public Health Accom- plishments Expected in the Coming Years: "(1) Sharp reduction in cancer death rates. (2) Reduction in the number of deaths and disabilities caused by high blood pressure. (3) Reduction in the infant mortality rate among minority groups." Why Did You )oin ACSH? "I was impressed with the goals of the Council in providing factual information on scientific and health issues and was and still'am~tired of the misinformation which is being pro- vided to the public by selected industrial groups and by the media, which exploit facts in order to make an issue 'sensa- tional.' "