Letters From the Global Literature: Unresolved Issues and Responses

Readers and Authors of a Controversial Recent JAMA Paper on Diet Debate Its Merits Here

Dean Ornish, MD; Barry Sears, PhD; John McDougal, MD; Caldwell B. Esselstyn, Jr., MD; T. Colin Campbell, PhD; Christopher D. Gardner, PhD; Randall Stafford, MD; Abby C. King, PhD


September 18, 2007

To the Editor:

A randomized controlled trial is designed to determine if differences between groups are due to an intervention or to random chance. When findings are not statistically significant, they are considered by convention to be due to random chance.

In this study, weight loss at 12 months was the stated primary outcome. The authors reported as their primary conclusion that people lost the most weight on the Atkins diet after 12 months. This was repeated in headlines worldwide.

However, this conclusion is not valid. The authors' own statistical analyses revealed that there was no statistically significant difference in weight loss among the Atkins, LEARN, and Ornish diets after 12 months. Thus, the small differences in weight loss after 1 year were most likely due to random chance. This is consistent with other studies finding that people who lost weight on an Atkins diet after 6 months gained it back after 1 year.[1]

Also, patients assigned to the Ornish diet did not follow it, reducing their fat intake from an average of 35% to only 30% after 1 year, not even close to the 10% fat recommended in the book. Other patients who followed the diet lost 24 pounds in the first year and maintained a 13-pound weight loss after 5 years.[2]

The authors also reported that patients had more favorable metabolic effects on an Atkins diet (eg, a significant rise in high-density lipoprotein cholesterol [HDL-C]). However, not everything that raises HDL-C is beneficial (as the Pfizer study of torcetrapib revealed), and not everything that lowers it is harmful. HDL-C is part of reverse cholesterol transport; eating a stick of butter raises HDL-C in those who are able to produce more of it, but this does not mean that butter is good for the heart.[3]

The authors did not mention randomized controlled trials showing that the diet and lifestyle intervention my colleagues and I recommend has been shown to reverse the progression of coronary heart disease (rather than only risk factors) using exercise radionuclide ventriculography,[4] cardiac PET scans,[5] and quantitative coronary arteriography,[2] as well as 2.5 times fewer cardiac events.[2] In contrast, coronary heart disease often progresses on an Atkins diet.[3]

The real conclusion of this study is that it's hard for many people to follow a diet just from reading a book and a few sessions with a dietitian. This is hardly news and very different from the reported conclusions.

Dean Ornish, MD
Sausalito, California


  1. Foster GD, Wyatt HR, Hill JO, et al. A randomized trial of a low-carbohydrate diet for obesity. N Engl J Med. 2003;348:2082-2090.

  2. Ornish D, Scherwitz L, Billings J, et al. Intensive lifestyle changes for reversal of coronary heart disease. Five-year follow-up of the Lifestyle Heart Trial. JAMA. 1998;280:2001-2007.

  3. Ornish D. Was Dr. Atkins right? J Am Diet Assoc. 2004;104:537-542.

  4. Ornish DM, Scherwitz LW, Doody RS, et al. Effects of stress management training and dietary changes in treating ischemic heart disease. JAMA. 1983;249:54-59.

  5. Gould KL, Ornish D, Scherwitz L, et al. Changes in myocardial perfusion abnormalities by positron emission tomography after long-term, intense risk factor modification. JAMA. 1995;274:894-901.



To the Editor:

The recent article by Gardner and colleagues[1] suffers from methodologic flaws that make their conclusions highly suspect. The 4 diets purported to be studied were not actually consumed by the participants as the protein, fat, and carbohydrate ratios did not reflect the stated dietary goals in the design. In particular, the absolute amounts of carbohydrate consumed were not consistent with the design of any of the diets. Specifically, the carbohydrate intake in the Atkins group was far too high, and the Zone group consumed too many carbohydrates relative to levels of protein. The LEARN group's carbohydrate intake was lower than recommended, while the Ornish group had too much total and saturated fats and was thus far too low in total carbohydrates. In fact, the amount of carbohydrates consumed on the Atkins group for the latter 6 months of the study was similar to my recommendations for the Zone Diet. In addition, the absolute amounts of carbohydrates consumed at the 12-month time point for the Ornish, LEARN, and purported "Zone" diets were very similar.

The authors also appear to be unaware of a published study in 2006 that compared a nonketogenic low-carbohydrate (Zone) diet to a ketogenic very-low-carbohydrate (Atkins) diet under tightly controlled conditions in which the meals for both groups were prepared exactly according the recommendations of both authors.[2] Although the weight loss and beneficial lipid changes were similar in both groups, inflammation levels doubled and vigor decreased in the Atkins group when compared with the Zone group.

In summary, since none of the diets was actually followed by the participants, and weight and lipid changes between the Zone diet and Atkins diet were the same in a more controlled clinical trial, the conclusions made by Gardner and colleagues[1] are most likely not valid.

Barry Sears, PhD
Inflammation Research Foundation
Danvers, Massachusetts


  1. Gardner CD, Kiazand A, Alhassan S, et al. Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in weight and related risk factors among overweight premenopausal women: the A TO Z Weight Loss Study: a randomized trial. JAMA. 2007;297:969-977.

  2. Johnston CJ, Tjonn SL, Swan PD, White A, Hutchins H, Sears B. Ketogenic low-carbohydrate diets have no metabolic advantage over nonketogenic low-carbohydrate diets. Am J Clin Nutr. 2006;83:1055-1061.



To the Editor:

Atkins Is an Easy Learning Curve

Whether someone follows a diet depends upon how easy it is to learn. The Atkins Diet is the easiest to follow -- you simply drive by a fast food window, order a burger, throw away the bun, and scrape off the pickles and ketchup, and you're on the diet. The Ornish Diet is much harder to learn. The foods are unfamiliar. They can take time and effort to prepare. The social stigma associated with being a vegetarian is daunting. With such a steep learning curve, few people succeed. This conclusion is substantiated by the observation that at 12 months, the group on the Ornish diet (a diet of 10% of the calories as fat) was actually consuming 29.8% fat. What we now need are studies that look at the long-term results for people who do follow various diets. Future investigations need to report the results of the effects of an Ornish type diet and the Atkins Diet on bowel movements. Research shows 70% of people following the Atkins Diet are constipated.[1] Anyone following the Ornish Diet knows the effects on bowel movements (often 3 times daily, easy to pass, and large). Next, investigators should look at calcium balance and see what happens to the bones on high-protein, high-acid diets like Atkins. Research consistently shows that a decrease in animal protein decreases loss of calcium.[2] Decreases in blood sugar, cholesterol, and triglycerides have been found with the Atkins Diet, but these changes are a result of suppression of appetite, followed by semi-starvation. Similar blood chemistry results can be accomplished by giving patients cancer chemotherapy. Rather than checking risk factors, like cholesterol and triglycerides, more direct measurements of the effect of diet on the heart and blood vessels need to be made. For example, a relevant measure would be the compliance of the artery walls. Results show a 27% decrease in arterial compliance after a single meal consisting of 67% of the calories as fat.[3] Reduction in blood flow in the heart arteries is also seen after one high-fat meal by use of a PET scan.[4] We need no more studies like this one to remind us of the fact that "diets fail" for most people -- but not for everyone. Let's study the effects when people do make meaningful changes.

John McDougal, MD


  1. Yancy WS Jr, Olsen MK, Guyton JR, Bakst RP, Westman EC. A low-carbohydrate, ketogenic diet versus a low-fat diet to treat obesity and hyperlipidemia: a randomized, controlled trial. Ann Intern Med. 2004;140:769-777.

  2. Giannini S, Nobile M, Sartori L, et al. Acute effects of moderate dietary protein restriction in patients with idiopathic hypercalciuria and calcium nephrolithiasis. Am J Clin Nutr. 1999;69:267-271.

  3. Nestel P. Post-prandial remnant lipids impair arterial compliance. J Am Coll Cardiol. 2001;37:1929-1935.

  4. Cook B, Cooper D, Fitzpatrick D, Smith S, Tierney D, Mehy S. The influence of a high fat meal compared to an Olestra meal on coronary artery endothelial dysfunction by rubidium (Rb)-82 positron emission tomography (PET) and on post prandial serum triglycerides. Clin Positron Imaging. 2000;3:150.



To the Editor:

The A TO Z weight loss study is a flawed attempt to evaluate different nutritional life styles. Weight loss at 12 months demonstrated no significant differences. The patients assigned to the Ornish group were not eating the Ornish dietary percentages of fat, protein, and carbohydrate. How can the authors make any claims from a short statistically insignificant study that did not achieve participant compliance? What are the truly long-term results?

In 1985 at the Cleveland Clinic, I initiated a plant-based study in 24 patients all with severe triple vessel coronary artery disease. I had not heard of Dr. Dean Ornish until 1987 when a patient showed me an article about him in Discovery Magazine. In response to my phone call and his gracious invitation, I visited his program. Our nutrition plans were similar. His patients performed meditation and relaxation. Mine did not. We agreed these were similar paths up the same mountain. Dr. Ornish's excellent results are well known.

I reported dramatically improved angina, lipid profiles, angiographic changes, and weight loss at 5 years.[1] At 12 years, the follow up was reviewed.[2] At 16 years, the compelling angiographic improvement was documented.[3] Finally, at 21 years, the results were again summarized, which makes it the longest arrest and reversal study of its type in the medical literature confirming that when patients are properly instructed about eating an Ornish-type plant-based diet, they maintain adherence and experience enduring vascular improvement along with sustained weight loss.[4] Particularly gratifying is a subset of patients who in 1985 were told by their expert cardiologists they had less than a year to live and are alive today, 21 years later. The Gardner study brings to mind an admonition of one of my surgical mentors: "Inappropriate application of the method is no excuse for its abandonment."

Contrary to the Atkins, LEARN, and Zone diets, the Ornish-type plant-based diet has a decades long track record of maintained weight loss, life extension, and arrest and reversal of vascular disease.

Caldwell B. Esselstyn, Jr., MD
Preventive Medicine Consultant, Cleveland Clinic
Cleveland, Ohio


  1. Esselstyn CB Jr, Ellis SG, Medendorp SV, et al. A strategy to arrest and reverse coronary artery disease; a 5 year longitudinal study of a single physician's practice. J Fam Pract. 1995;41:560-568.

  2. Esselstyn CB Jr. Updating a 12 year experience with arrest and reversal therapy for coronary heart disease. Am J Cardiology. 1999;84:339-341.

  3. Esselstyn CB Jr, Resolving the coronary artery disease epidemic through plant based nutrition. Prev Cardiol. 2001;4:171-177.

  4. Esselstyn CB Jr. Prevent and Reverse Heart Disease. New York, NY: Avery; 2007.



To the Editor:

Gardner and colleagues[1] failed to test the impressive ability of diet to favorably affect long-term health and body weight loss. The authors' suggestion that their study represents a "substantial" spectrum of carbohydrate intake is a superficial view. The Atkins and Zone diets are unusually high in protein and fat and low in carbohydrate while the LEARN diet is only slightly different from the high-protein, high-fat American diet that is associated with high rates of degenerative diseases typically found in industrialized societies.[2]

The authors' claim that the Ornish diet is "extremely low in fat" misrepresents. At 12 months, the Ornish diet being used included 30% fat calories, 18% protein calories, and only 20 g/day of dietary fiber. This is not the original Ornish diet[3] or the plant-based, whole foods diet (with little or no added fat or refined carbohydrate) that remarkably cures heart disease[3,4] and associates with 110-130 mg/dL blood cholesterol and extremely low rates of diseases[5] typically found in Western countries. The Ornish diet in this study includes 48% fat plus protein calories whereas a typical whole foods, plant-based diet with no added fat, animal foods, and refined carbohydrates contains about 20% fat plus protein and a minimum of 30-40 g/day of dietary fiber (instead of the 20 g/day used here). This is a huge difference not represented in this study. Diets containing animal protein in excess of the recommended 10% protein, along with its usual correlates, present myriad adverse health effects, including increased circulating cholesterol, increased atherogenesis, and increased cancer and other degenerative diseases, as reviewed elsewhere.[5]

All 4 diets used in this study represent nutrient excesses that can only be expected to produce trivial and confusing differences in health outcomes. It's tantamount to predicting long-term health consequences of tobacco use by comparing 3 packs/day cigarette smokers with 4 packs/day smokers 12 months after starting the habit. Finally, the trivial "benefits" claimed for the Atkins diet in this very short study should be considered in relation to the long-term adverse health outcomes noted for diets high in fat, cholesterol, animal protein, and processed foods when consumed for a lifetime. The suspension of the initial weight lost during the first 6 months, the unspecified attrition rate and adverse effects for this diet commonly observed in other studies,[6] and the trivial increase in low-density lipoprotein cholesterol point to the emergence of these long-term adversities.

T. Colin Campbell, PhD
Jacob Gould Schurman Professor Emeritus of Nutritional Biochemistry, Cornell University, Ithaca, NY
Lansing, NY


  1. Gardner CD, Kiazand A, Alhassan S, et al. Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in weight and related risk factors among overweight premenopausal women: the A TO Z Weight Loss Study: a randomized trial. JAMA. 2007;297:969-977.

  2. US Department of Health and Human Services. The Surgeon General's Report on Nutrition and Health. Washington, DC: Superintendent of Documents, US Government Printing Office; 1988.

  3. Ornish D, Brown SE, Scherwitz LW, et al. Can lifestyle changes reverse coronary heart disease? Lancet. 1990;336:129-133.

  4. Esselstyn CJ. Prevent and Reverse Heart Disease. New York, NY: Avery; 2007.

  5. Campbell TC, Campbell TM II. The China Study, Startling Implications for Diet, Weight Loss, and Long-Term Health. Dallas, Tx: BenBella Books Inc; 2005.

  6. McDougall J. The Atkins scientific research--deceit and disappointment. McDougall Newsletter. 2004;3:1-17.



Response to Letters to the Editor:

The A TO Z study was designed to test the 12-month effectiveness of popular weight loss diets among 311 overweight or obese women. By design, the primary intervention was to read and follow assigned diet books.[1] The specific diet books selected were chosen primarily on the basis of their strikingly different recommendations for carbohydrate intake -- from the very-low-carbohydrate Atkins diet to the very-high-carbohydrate Ornish diet, with the Zone and LEARN diets being intermediate. After 12 months, the group assigned to the Atkins diet lost an average 4.7 kg while the other 3 groups lost averages of 1.6-2.6 kg. Drs. Campbell, MacDougal, Sears, Esselstyn, and Ornish suggest that poor adherence may have led to inadequate testing of the diets. Documenting adherence was a major priority, as evidenced by the resources committed to collecting 3137 dietitian-administered 24-hour recalls. However, attempting to enforce adherence, as one might do in a feeding study,[2] or other laboratory experiment, was not a study aim. Rather, the primary objective was to test the effectiveness of popular weight loss diet books followed under free-living or "real world" conditions. At 12 months, there were statistically significant group differences in macronutrient intakes and weight loss. Adherence was mediocre, reflecting real-world challenges with each diet book, despite optimization of program delivery via dietitian assistance in the first 8 weeks of the protocol. The development of methods to optimize adherence to weight loss diets through increased use of behavior modification, improved physical activity strategies, and other approaches is a research priority.[3]

Drs. Sears, Esselstyn, and Ornish each cite some of their own work for the purpose of comparing results with our trial. Dr. Sears has reported findings from a 6-week feeding study of 20 overweight adults.[2] Drs. Esselstyn[4] and Ornish[5] have reported their findings from working intensively with small numbers of adults (24 and 28 patients, respectively) with severe coronary artery disease. Although some of the types of diets used in these small, intensive studies and the A TO Z study were the same, the study objectives, designs, and populations were so different as to make it inappropriate to directly compare their results.

The A TO Z study addressed 12-month weight loss as the primary outcome. Secondary outcome measures included blood lipids, blood pressure, insulin, and glucose. We are in agreement with the comments suggesting that health outcomes such as bone density, renal function, and cardiovascular events are important in assessing the overall health impact of weight loss diets. However, these are outcomes that require follow-up over periods of time longer than 12 months and were beyond the scope of our study. Large prospective observational studies are important sources of information for these questions. Findings from the Nurses Health Study indicate that lower vs higher carbohydrate diets were associated with decreased coronary heart disease, but only when fat and protein sources were derived primarily from vegetable rather than animal origin.[6] Dr. Campbell has reported associations between diets of low animal protein content and favorable morbidity and mortality rates in China.[7] Both of these observational studies are important contributions to our knowledge of diet patterns and health, but neither of these studies were designed to specifically address weight loss, and neither can address causality, only associations.

In the A TO Z trial, at 12 months, all 4 groups had lost modest amounts of weight relative to baseline and experienced corresponding improvements in most of the metabolic variables that were measured as secondary study outcomes, including lipids, blood pressure, and insulin. Primary and secondary diet group differences at 12 months were analyzed using intention-to-treat principals (baseline values carried forward for drop outs) and included adjusting for multiple comparisons among groups. At 12 months, there were no statistically significant differences among 3 of the 4 diet groups: Zone, LEARN, or Ornish. In contrast, changes in several variables were statistically significant for the Atkins vs other groups: weight loss and triglyceride lowering was greater for the Atkins than the Zone group, blood pressure reductions were greater for the Atkins group than all 3 other groups, and HDL-C changes were more favorable for the Atkins vs the Ornish group. In other words, all the statistically significant findings in the study favored the Atkins group. However, in each case, the clinical significance of the group differences was relatively modest. Based on these findings, our general conclusion from the study is that a very low carbohydrate diet should be considered a feasible alternative weight loss approach.

We do not suggest that these findings support abandoning all other dietary weight loss approaches in favor of a very low carbohydrate diet. But at the same time, these findings do not support the more traditional "low fat/high carbohydrate" diet as being generally superior to low carb diets for weight loss. Given that a wide range of different macronutrient distributions can be used to achieve weight loss in general, the next important question is whether there are individual characteristics that could be used to identify certain people who would be more or less likely to succeed with weight loss on certain diets. We are currently examining our data for these potentially useful identifiers. We also strongly believe that further research is warranted to examine mechanistic explanations for the effects of very low carbohydrate diets and to examine the long-term, overall health consequences of different dietary weight loss approaches.

Christopher D. Gardner, PhD
Randall Stafford, MD
Abby C. King, PhD
Stanford, California


  1. Gardner CD, Kiazand A, Alhassan S, et al. Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in weight and related risk factors among overweight premenopausal women: the A TO Z Weight Loss Study: a randomized trial. JAMA. 2007;297:969-977.

  2. Johnston CS, Tjonn SL, Swan PD, White A, Hutchins H, Sears B. Ketogenic low-carbohydrate diets have no metabolic advantage over nonketogenic low-carbohydrate diets. Am J Clin Nutr. 2006;83:1055-1061.

  3. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults -- The Evidence Report. National Institutes of Health. Obes Res. 1998;6(suppl 2):51S-209S.

  4. Esselstyn CB Jr. Updating a 12-year experience with arrest and reversal therapy for coronary heart disease (an overdue requiem for palliative cardiology). Am J Cardiol. 1999;84:339-341, A338.

  5. Ornish D, Brown SE, Scherwitz LW, et al. Can lifestyle changes reverse coronary heart disease? The Lifestyle Heart Trial [see comments]. Lancet. 1990;336:129-133.

  6. Halton TL, Willett WC, Liu S, et al. Low-carbohydrate-diet score and the risk of coronary heart disease in women. N Engl J Med. 2006;355:1991-2002.

  7. Campbell TC, Campbell TM. The China Study: The Most Comprehensive Study of Nutrition Ever Conducted and the Startling Implications for Diet, Weight Loss and Long-term Health. Dallas, Tx: BenBella Books; 2005.


Reader Comments on: Letters From the Global Literature: Unresolved Issues and Responses: Readers and Authors of a Controversial Recent JAMA Paper on Diet Debate Its Merits Here
See reader comments on this article and provide your own.

Readers are encouraged to respond to Paul Blumenthal, MD, Deputy Editor of MedGenMed, for the editor's eyes only or for possible publication as an actual Letter in MedGenMed via email:


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.