Diet and cancer: a damning analysis

This month’s edition of the American Journal of Clinical Nutrition carries a very important review of the epidemiological data relating food to cancer[1]. It is well written and sensitive in its conclusions but, reading between the lines, it is quite simply damning of the quality of the epidemiological research basis linking food intakes and cancer. The authors started off with The Boston Cooking-School Cook Book.  Using a random number generator to correspond to page numbers, they searched the cook-book for recipes. All of the unique ingredients in each recipe was identified and the process was repeated until 50 unique food ingredients were identified. The next stage of the process was to explore the scientific literature to examine the most recent studies, if any, linking any one of the 50 ingredients to cancer. The 10 most recent studies were selected and if there were less than 10 studies available, synonyms (e.g. mutton for lamb) were used to further explore the availability of studies. In addition to individual studies, the authors also searched for meta-analysis studies that combine data from several individual studies to increase statistical power.

 

The next stage of the process was to extract data from each individual study or meta-analysis. This involved an examination of the abstract with an emphasis on the author’s conclusions, an analysis of the statistical methodology used and an assessment of the exposure levels examined for each ingredient. From the 50 ingredients randomly chosen from the cookbook, 40 (80%) were found to be the subject of a scientific investigation into its links with cancer. The food ingredients included: veal, salt, pepper spice, flour, egg, bread, pork, butter, tomato, lemon, duck, onion, celery, carrot, parsley, mace, sherry, olive, mushroom, tripe, milk, cheese, coffee, bacon, sugar, lobster, potato, beef, lamb, mustard, nuts, wine, peas, corn, cinnamon, cayenne, orange, tea, rum and raisin. A total of 216 publications were found linking these ingredients to cancer. Of the 40 ingredients, 36 were identified in at least one study as either increasing or decreasing the risk of cancer. In their examination of the statistical analysis used, the authors of the review concluded that, of the studies that claimed an increased risk of cancer, 75% were associated with “weak or non-nominally significant” effects and for those that reported a negative effect, the corresponding figure was 76%.  In 65% of the studies, these effects were based on comparison of extremes of consumption such as >43 drinks per week versus none or “often” compared with “never”.  The authors compared the calculated risk from individual studies with the risk calculated from meta-analyses where like studies were pooled. The latter showed a narrow range of risk where as the former showed a huge range of positive and negative risk.  

The authors conclude that the vast majority of claims for increased or decreased risk were based on “weak statistical evidence”. Moreover, they show an appalling practice of relegating negative or weak results to the fine detail of the text of the paper but excluding such from the abstract. The abstract is most likely to be read and certainly to be the basis of any media interpretation of the study.

All in all, this is a damning analysis of the field of nutritional epidemiology. It would be wrong to throw the baby out with the bathwater since nutritional epidemiology has been the basis for many substantiated diet-disease links (folic acid and neural tube defects, atherosclerosis and dietary lipids, calcium intake and osteoporosis and so on).  The problem with diet and cancer is that intervention studies to experimentally prove a cause and effect relationship are simply not possible. Heart disease relates to one organ, the heart, whereas cancer can relate to almost all organs. In the study of diet and heart disease we can use biomarkers (plasma HDL and LDL cholesterol for example) while no such biomarkers exist for cancer.

In the analysis of how extremes of exposure were used, the authors found that the meta-analysis approach was almost exclusively based (92%) on extremes of consumption of the particular food ingredient. Epidemiological studies, by definition are very large and as such, the tools to measure diet must be relatively simple and this usually involves a food frequency questionnaire, which examines frequency of intake. However, nutritional epidemiology absolutely ignores the well know phenomenon of under-reporting of food intake.  Thus, when extremes of food intake are compared, they are largely based on truly flawed measures of food intake. This was not considered in the present study and more often than not, insufficient data is presented in papers to allow readers to make any conclusions as to the extent of under-reporting based on the match of reported energy intake and estimated energy requirements. The editorial boards of journals should insist that all studies involving food intake have a section in which the authors explain the extent of energy under-reporting and the specific implications for the study in question.

No doubt, this damning review of the epidemiology of diet and cancer will be swept under the carpet by the field of nutritional epidemiology. However, this blogger has always held the view that bad science will always be found out.

---

[1] Schoenfeld JD & Ioannidis JPA (2013) “Is everything we eat associated with cancer? A systematic cookbook review” Am J Clin Nutr 97, 127

Treating diabetes: Diets, gyms and scalpels

In the US, 27% of those aged 65 or older have diabetes. Based on fasting blood glucose levels or glycated haemoglobin levels, the estimate of the prevalence of pre-diabetes is 79 million cases. The economic cost has been estimated at around $2,000 per person per annum. The maths aren’t complicated working out at about $22 billion per annum.  Most  cases of type 2 diabetes are treated initially by lifestyle changes and then by drugs to manage blood glucose. In 2009, the American Diabetes Association defined partial remission from diabetes when fasting blood glucose levels were lowered to below the diagnostic norm and complete remission when fasting blood glucose levels returned to normal, in both cases in the absence of drug therapy. Relatively little is really known of the extent to which such partial or complete remission can be achieved with lifestyle interventions. In 2001, a large multicenter study was established in the US known as “Look Ahead”, (Actions for Health in Diabetes)^[1]. The trial, funded by the National Institute of Health, involved 2,262 type 2 diabetics given a basic diabetes lifestyle intervention and a group of 2,241 type 2 diabetics given an intensive lifestyle intervention. The former were given 3 group sessions per year while the latter participated in weekly group and one-on-one counseling for the first 6 months followed by 3 sessions per month for the next 6 months and twice monthly sessions thereafter. For this group, a target caloric intake was set between 1,200 and 1,800 calories per day with an exercise goal of 175 minutes of moderately intensive exercise per week (25 minutes per day).  The Look Ahead trial laid out its hypothesis quite clearly: that there would be a significant reduction in heart disease and stroke in the intensively counseled group compared to the group receiving standard advice on lifestyle. The trial has produced over 80 peer reviewed papers and has shown that intensive lifestyle intervention can significantly improve body-weight, blood pressure, blood glucose control and blood lipid levels.

On October 19th this year, when the trial was well into its 11th year, the NIH announced the end of the trial on foot of recommendations from the trial’s data and safety monitoring board. This independent body of experts noted that despite the above improvements on risk factors for cardiovascular disease, there was no statistically significant difference in cardiovascular events between the two groups which was the central hypothesis. Recently, the trial study group published a paper in the Journal of the American Medical Association showing that intensive lifestyle intervention did indeed lead to a greater rate of remission of type 3 diabetes compared to the standard intervention^[2]. The big disappointment was, however, that the impact of intensive lifestyle was very small. The rate of partial or complete remission in year 1 was 11. 5% in the intensively tutored group, falling to 7.3% at year 4. In contrast, the group receiving standard counseling showed a 2% reduction at both time points. Very clearly, type 2 diabetes is not a reversible condition for the vast majority of subjects. And just as clearly, this low response rate in correcting diabetes pathologies explains why no differences in heart disease were observed between the two treatment groups.

In the same issue of this journal, an editorial looks at the overall evidence for lifestyle and surgical interventions in obesity^[3]. The latter are usually confined to subjects with very severe cases of obesity. The latter leads to type 2 diabetes remission rates, which are 12 to 24 fold greater than intensive lifestyle interventions. The Swedish Obesity Study, also published in this year’s JAMA, reported on the long-term effects of the surgical treatment of obesity. Subjects were morbidly obese at baseline (1987 was the start date) and the average duration of follow up was 14.7 years^[4]. Compared to conventional medical and lifestyle treatment, the surgical intervention reduced fatal heart attacks by 47%, all heart attacks by 52% and stroke by 34%. Surgery is expensive but so too is intensive lifestyle interventions and thus some cost comparisons between the two would be interesting.

Clearly, we are in a mess and we must now live with the mess. But how can we prevent the mess for future generations?. Whilst 79 million Americans have prediabetes and are at risk of developing diabetes, the remaining 233 million don’t. Of those aged over 65 years, 11.2 million have type 2 diabetes while the remaining 30 million over 65s do not. They all live in the same obesogenic US environment. One day, not far from now, we will be able to predict who is likely to to draw the short straw and develop obesity-related type 2 diabetes. Moreover, this genetic information will soon be able to zone in on that aspect of diet and lifestyle, which is most responsible for the development of diabetes. For some, it may be a metabolically based genetic factor. For others, it may be a food choice factor that is the driver and for others it may be a defective satiety system. Understanding personal risk and understanding personalised solutions is the future for nutrition and health. In the meantime, we have a mess.

---

^[1] https://www.lookaheadtrial.org

^[2] Gregg et al (2012) JAMA 308,2489

^[3] Arterburn DE & O’Connor PJ (2012) JAMA 308, 2517

^[4] Sjostrom L et al (2012) JAMA 307, 56

Obesity and Nature v Nurture re-visited

In the obesogenic environment that we live in, not everyone becomes obese. To the high priests of nutrition, that variability is put down to variation in self-control and self-discipline and that in turn relates to level of education and social class. The idea that this variation might be genetically based is dismissed with the old reliable falsism that since our genes have not changed during the recent epidemic of obesity, it’s the environment that counts. Well, yet another twin study shows that this is nonsense and this twin study is somewhat special since it pooled data from 23 twin cohorts from four countries: Denmark, Australia, Canada and Sweden involving just over 24,000 children[1]. Moreover, this pooling study was able to provide data on twins from birth through 19 years of age. By comparing variation within and between both identical and non-identical twins, it is possible to distinguish the effect of genes from the effect of the environment and the latter can be split into common and unique environments. At birth, only 8% of variation in weight or body mass index (BMI) could be explained by genetic factors. By 5 months this had increased to 65% and rose into the 70% decile up to 9 years of age. In the early teens the genetic variation had reached into to 80% decile and by late teens it had hit 90%.  As children got older, the environmental explanation of obesity had fallen from 74% at birth, to 25% at 6 years and down to about 10% in late teens. While this study clearly shows the powerful effect of genetic factors on obesity, it does raise the question as to why this genetic dimension increased with age. Clearly, the genetic make up remained constant so most likely, changes in gene expression were the contributory factor. Growth in childhood and especially in adolescence is associated with significant biological adjustments, which could create the environment for altered gene expression.

One of the reasons which I personally think public health nutritionists are wary of the genetic influence on obesity is that the subject is strongly orientated toward basic biology, effectively, the digestion, absorption, transport, distribution and utilisation of calories from fat, carbohydrate, protein and alcohol. However, genetic influences on behaviour are to my mind far more important   than the genetics of basic biological elements. A recent twin study has looked at the heritability of taste[2].  Subjects were given a strawberry jelly with or without the hot spice capsaicin derived from chili peppers. They were also asked questions on their liking or otherwise of spicy foods and spices and of foods that have mild, strong and extremely strong pungency properties. 50% of the variation in preference for spicy foods and spices and 58% of the variation in “pleasantness of strong pungency” was explained by genetic factors. Another twin study looked at food neophobia in a group of children aged 8 to 11 years, comprising 5,390 pairs of identical and non-identical twins[3].  Parents were asked about their children’s attitude to foods with four statements: “My child is constantly sampling new and different foods”, “My child doesn’t trust new foods,” “My child is afraid to eat things/he has never had before.” and “If my child doesn’t know what’s in a food s/he won’t try it.” A food neophobia score was worked out and the highly robust finding of the study was that a staggering 78% of variation in food neophobia was genetic in origin. Only 22% was learned from the environment. These studies show that the genetic component of obesity need not be related to the biochemistry of energy metabolism, but rather to more complex behavioural traits such as food choice.

Twin studies of obesity always raise the question of assortative mating, that is fat partners mating with other fat partners and similarly for slim partners. Assortative mating has been shown to occur in personality type, education, religion, politics, age, smoking habits and anti-social behaviour. Researchers at the Rowett Institute in Aberdeen used DEXA scans to accurately measure body fat levels in 42 couples[4]. Strong evidence for assortative mating in relation to body fat was found. For example, subjects with disproportionately large arms assortatively mated with like partners. Given the high heritability of the propensity to develop obesity, assortative mating will accelerate the incidence of obesity sine the children of such parents are likely to inherit genetic patterns from both parents.  

The high priests of public health nutrition may dislike the implications of a genetic dimension to obesity but they are being increasingly isolated from the scientific truth.

---

[1] Dubois et al (2012) PLoS ONE 7, e30153

[2] Tornwall et al (2012) Physiology & Behaviour, 107, 381-389

[3] Cooke et al (2007) Am J Clin Nutr 86, 428-433

[4] Speakman et al(2007) Am J Clin Nutr 86, 316-323

Sugar taxes and weight loss predictions

The Danish government has abandoned its tax on fat and and its plans for a sugar tax.  A spokesperson for the tax ministry is quoted thus: “The suggestions to tax foods for public health reasons are misguided at best and may be counter-productive at worst.  Not only do such taxes not work, especially when they choose the wrong food to tax, they can become expensive liabilities for the businesses forced to become tax collectors on the governments behalf”[1]. Shortly we will have our annual budget here in Ireland and notwithstanding the volte-face of our Danish colleagues, the likelihood is that we will face such a tax soon.  In general, the predicted weight changes associated with projected taxes on sugar sweetened beverages are grossly overestimated.

A recent consensus statement of the American Society of Nutrition (ASN) and the International Life Sciences Institute (ILSI) has examined the topic:  “Energy balance and its components:  Implications for body weight regulation”[2].  One of the areas covered by this paper is the popular and widely held belief that to lose 1lb of body weight, you need to reduce caloric intake by 3,500 kcal.   This figure assumes that a loss of 1lb of body weight is made up entirely of adipose tissue which is 86% fat and the fat has 9 kcal per gram.  This 3,500 kcal figure is widely used in predicting the benefit of weight loss from a sugar sweetened beverage tax.  It has many flaws.

Firstly, a 1lb weight loss will not be 100% fat but will also involve the loss of some lean tissue (muscle and protein elements of adipose tissue and its metabolism).  Whereas fat has an energy value of 9 kcal/g, lean tissue has a value of 4 kcal/g.  The exact ratio of the loss of lean and fat in weight reduction depends largely on the level of fat in the body at the outset.  The higher the intake level of fat, the higher the proportion of fat lost.  However, as a person sheds fat, the ratio of fat to lean changes in favour of the latter, so subsequent weight loss will have a lower ratio of fat to lean.  The blanket use of the 3,500kcal value ignores this.

The second criticism of this rule is that it ignores time.  If you shed 3,500kcal per week every week, that would differ from a deficit of 3,500 kcal per month every month.  The former leads to a daily deficit of 500 kcal while the latter is just 117 kcal.  Even the most non-expert dieter knows that such differences in daily energy deficits will lead to radically different rates of weight loss.  Thirdly, the 3,500 kcal rule assumes complete linearity – in other words the rule equally applies, pound after pound of weight loss. We saw above that progressive weight loss will progressively increase the % of that weight loss as lean tissue but more importantly, the 3,500kcal rule ignores a major adaptation in energy expenditure.  Basically, our basal metabolic rate (BMR) falls as we restrict our caloric intake.  Since BMR accounts for 88% of energy expenditure in most sedentary persons, that means that a fall in BMR represents a significant adaptive response through increased efficiency of energy use making weight loss progressively more difficult.

Researchers at the US National Institute of Health have developed a very detailed mathematical model which predicts weight loss based on a wide variety of inputs[3].  The model has been validated against a number of highly controlled weight loss programmes.  Together with researchers based at the USDA and the economics departments of the universities of Florida and Minnesota, they have examined the likely weight loss that would accrue from a tax of 20% (about 0.5 cents per ounce) on sugar sweetened beverages in the US[4].  They concluded that the nutritional input would be a reduction of energy intake of 34-47 kcal per day for adults.  Using the 3,500 kcal rule, an average weight loss of 1.60kg would be predicted for year 1 rising to 8kg in year 5 and to 16kg in year 10.   However, when the dynamic mathematical model is used, the corresponding figures for years 1, 5 and 10 are, respectively, 0.97, 1.78 and 1.84 kg loss.  The % of US citizens that are over-weight is predicted to fall from existing levels of 66.9% over-weight to 51.5% over-weight in 5 years time using the 3,500 kcal rate but using the dynamic mathematical model, the 5-year figure for the over-weight population in the US would be just 62.3%.  Clearly, the continued use of the 3,500 kcal rule in predicting weight loss should cease and the recommendations of the consensus statement of the ASN and ILSI should apply: “Every permanent 10 kcal change in energy intake per day will lead to an eventual weight change of 1lb when the body reaches a new steady state.  It will take nearly a year to achieve 50% and about 3 years to achieve 95%”.

My back of envelope calculations based on the National Adult Nutrition Survey is that extrapolating from the US model (footnote 4), a tax on sugar sweetened beverages might lead to a weight loss of 0.6 lb at the end of year 1. That of course is subject to an error estimate such that it might be higher but equally, it might be lower. Many of the advocates of fat taxes might argue that they will take that “thank you very much” as a start and then move to the next food. But you cannot continue to add tax to the cost of food.

---

[1] http://www.foodnavigator.com/Legislation/Danish-government-scraps-fat-tax-cancels-planned-sugar-tax

[2] Hall et al (2012) Am J Clin Nutr 12, 989-994

[3] Hall et al (2011) Lancet 378,826-837

[4] Biing-Hwan et al (2011) Econ Hum Biol 9, 329-341