To the Editor: How do Shai et al. (July 17 issue)1 explain whythe subjects in their study regained weight between month 6and month 24, despite a reported reduction of 300 to 600 caloriesper day? Contributing possibilities may include the notion thata food-frequency questionnaire cannot precisely determine energyor macronutrient intake but, rather, ascertains general dietarypatterns. Certain populations may underreport intake2,3 andhave a decreased metabolic rate. The authors did not measurebody composition, which is critical for documenting weight-losscomponents.
In addition, the titles of the diets that are described in thearticle are misleading. Labeling the "low-carbohydrate" dietas such is questionable, since 40 to 42% of calories were fromcarbohydrates from month 6 to month 24, and data regarding ketosissupport this view. Participants in the low-fat and Mediterranean-dietgroups consumed between 30% and 33% of calories from fat anddid not increase fiber consumption, highlighting the importanceof diet quality. Furthermore, the authors should have providedbaseline values and P values for within-group changes from baseline(see Table 2 of the article).
Contrary to the authors' assertion, it is not surprising thatthe effects on many biomarkers were minimal, since the dietarychanges were minimal. The absence of biologically significantweight loss (2 to 4% after 2 years) highlights the fact thatenergy restriction and weight loss in themselves may minimallyaffect metabolic outcomes and that lifestyle changes must incorporatephysical activity to optimize the reduction in the risk of chronicdisease.4,5
Christian K. Roberts, Ph.D. R. James Barnard, Ph.D. Daniel M. Croymans, B.S. University of California, Los Angeles Los Angeles, CA 90095-1606 croberts{at}ucla.edu
References
Shai I, Schwarzfuchs D, Henkin Y, et al. Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet. N Engl J Med 2008;359:229-241. [Free Full Text]
Schoeller DA. Limitations in the assessment of dietary energy intake by self-report. Metabolism 1995;44:Suppl 2:18-22. [ISI][Medline]
Neuhouser ML, Tinker L, Shaw PA, et al. Use of recovery biomarkers to calibrate nutrient consumption self-reports in the Women's Health Initiative. Am J Epidemiol 2008;167:1247-1259. [Free Full Text]
Roberts CK, Barnard RJ. Effects of exercise and diet on chronic disease. J Appl Physiol 2005;98:3-30. [Free Full Text]
Booth FW, Chakravarthy MV, Gordon SE, Spangenburg EE. Waging war on physical inactivity: using modern molecular ammunition against an ancient enemy. J Appl Physiol 2002;93:3-30. [Free Full Text]
To the Editor: Shai and colleagues found that a low-fat dietwas less effective than either a Mediterranean diet or a low-carbohydratediet in achieving weight loss. However, the "low-fat" groupwas actually not instructed to follow a low-fat diet, becausethe diet composition (i.e., calories from fat) did not changeduring the intervention. So, in reality, since the members ofthis group consumed the same diet as they did habitually, itshould have been called the control group. Weight loss is knownto occur as a consequence of reducing the proportion of caloriesfrom fat,1 so the design of the study was not really fair tothe low-fat diet. Moreover, the dropout rate was 15% in theMediterranean group, 22% in the low-carbohydrate group, andonly 10% in the low-fat group (P=0.04). In a study involvingDanish subjects,2 my colleagues and I also observed that therewas a significantly higher dropout rate among participants ona Mediterranean diet (28%) than among those on a (real) low-fatdiet (16%), after both 6 months and 18 months. Retention isan end point that deserves comment.
Arne Astrup, M.D. University of Copenhagen 1958 Frederiksberg, Denmark ast{at}life.ku.dk
Dr. Astrup reports receiving consulting fees from Weight Watchersand Global Dairy Platform and reports that his department atthe University of Copenhagen has received research funding frommany food companies. No other potential conflict of interestrelevant to this letter was reported.
References
Astrup A, Grunwald GK, Melanson EL, Saris WHM, Hill JO. The role of low-fat diets in body weight control: a meta-analysis of ad libitum intervention studies. Int J Obes Relat Metab Disord 2000;24:1545-1552. [CrossRef][ISI][Medline]
Due A, Larsen TM, Mu H, et al. Comparison of three ad libitum diets for weight loss maintenance, risk for CVD and diabetes: a 6 month randomised, controlled trial. Am J Clin Nutr (in press).
To the Editor: Shai et al. report that either a Mediterraneandiet or a low-carbohydrate diet was an effective alternativeto a low-fat diet in terms of weight loss. Absolute values forthe energy intake before and after the intervention are notgiven for any of the groups. However, the data provided in thearticle suggest that the weight loss in the low-carbohydrategroup may have been simply a hypocaloric effect, even thoughthe authors state that the energy intake in this group was notrestricted. Thus, if the daily intake of carbohydrates was nomore than 120 g (equal to 500 kcal), as stated in the Methodssection, and if the fraction of the energy intake provided bycarbohydrates was around 40%, as stated in Table 2, then thetotal daily energy intake in the low-carbohydrate group musthave been no more than approximately 1250 kcal, which is significantlyless than the daily intake targeted for the two other groups.
To the Editor: In the study by Shai et al., participants whowere on the low-fat diet decreased their total fat intake from31.4% to 30.0% — in other words, hardly at all. It isimportant to measure disease end points, not just risk factors.Greater increases in high-density lipoprotein (HDL) cholesterolamong patients on the Atkins diet may not necessarily be beneficial,as the torcetrapib study indicated.1 Very-low-fat diets mayreverse coronary heart disease2 and prostate cancer3 and alsoimprove gene expression,4 despite reductions in HDL cholesterol,whereas flow-mediated vasodilatation, a measure of endothelialfunction and a predictor of cardiovascular events, worsenedamong subjects on the Atkins diet.5
In the "Atkins diet" that was tested by Shai et al., the authorsstate that they counseled the study participants to choose vegetariansources of fat and protein and to avoid trans fat. However,such a diet is not representative of the Atkins diet, whichrecommends bacon, butter, and brie, not fruits and vegetables.Why were the participants counseled to do this?
Dean Ornish, M.D. University of California, San Francisco San Francisco, CA 94143
Dr. Ornish reports writing general-interest books on healthand wellness (including diet) for which he receives royalties.He also reports receiving lecture fees from the Harry Walkerspeaker's bureau and consulting fees from PepsiCo, Mars, andSafeway. No other potential conflict of interest relevant tothis letter was reported.
References
Barter PJ, Caulfield M, Eriksson M, et al. Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med 2007;357:2109-2122. [Free Full Text]
Ornish D, Scherwitz LW, Billings JH, et al. Intensive lifestyle changes for reversal of coronary heart disease. JAMA 1998;280:2001-2007. [Erratum, JAMA 1999;281:1380.] [Free Full Text]
Ornish D, Weidner G, Fair WR, et al. Intensive lifestyle changes may affect the progression of prostate cancer. J Urol 2005;174:1065-1070. [CrossRef][ISI][Medline]
Ornish D, Magbanua MJM, Weidner G, et al. Changes in prostate gene expression in men undergoing an intensive nutrition and lifestyle intervention. Proc Natl Acad Sci U S A 2008;105:8369-8374. [Free Full Text]
Miller M, Beach V, Mangano C, et al. Comparative effects of 3 popular diets on lipids, endothelial function and biomarkers of atherothrombosis in the absence of weight loss. Circulation 2007;116:II-819.
To the Editor: Shai et al. report very low attrition rates amongstudy participants, and this is surely a good point, given the2-year length of the study. The authors properly recognize thatthis finding is probably due to a closely monitored intervention,but they do not report anything about the costs of their face-to-face,dietitian-based approach. As noted by Glasgow and Emmons,1 amongfactors that impede the translation of research into widespreadpractice, important obstacles are the high cost and large timedemands on both staff and participants. Even if the authors'strategies for maintaining adherence could be applied outsidethe workplace, we are not as optimistic as they are, consideringthe low cost-effectiveness that usually characterizes face-to-faceapproaches targeting eating habits.
Gian Mauro Manzoni, Psy.D. Gianluca Castelnuovo, Ph.D. EnricoMolinari, Ph.D. Istituto Auxologico Italiano 28824 Piancavallo, Italy gm.manzoni{at}auxologico.it
References
Glasgow RE, Emmons KM. How can we increase translation of research to practice? Types of evidence needed. Annu Rev Public Health 2007;28:413-433. [CrossRef][ISI][Medline]
The authors reply: Roberts and colleagues correctly note thata food-frequency questionnaire cannot precisely assess exactcaloric intake. In our study, we used identical nutritionalvalues for each of the 137 questionnaire items, irrespectiveof the specific dietary intervention. For example, "vegetablesoups" were not specially analyzed within the low-fat groupas "no fat added" or in the low-carbohydrate group as "creamed"or "no starch added." Thus, although the questionnaire datarevealed three distinct dietary regimens throughout the intervention,these data might have underestimated the true differences amongthe groups. Correlations between data from food-frequency questionnairesand those from 24-hour dietary recalls, which were used in asubgroup of participants, were more than 0.8 for most nutrients.
In response to the letters from Astrup and Moller and Krogh-Madsen:the diet-recall data confirm that the total caloric deficitwas similar among the groups and that the low-fat group maintaineda relatively low intake of fat (Table 1). Obviously, with asubstantial total caloric deficit, the absolute consumptionof fiber was decreased; however, the decrease in the low-fatgroup was half of that in the low-carbohydrate group.
Table 1. Dietary Intake from 24-Hour Dietary Recall among Participants in the Dietary Intervention Randomized Controlled Trial (DIRECT).
We agree with Roberts et al. on the importance of physical activity,but our study specifically focused on diet composition. Robertset al. are incorrect in characterizing the weight loss as biologicallynot meaningful, since it yielded clinically important changesin risk factors. Intriguingly, the weight reduction achievedamong the moderately obese subjects in our study was identicalto that reported in a recent meta-analysis on long-term pharmacotherapyfor obesity.2
In response to Ornish's letter: the low-carbohydrate diet wasbased on Atkins's book, although we encouraged the study participantsto consume a variety of protein and fat sources to maintainsuccessful long-term adherence. We specifically reported onretention in detail and noted that adherence was higher in thistrial than in any other trial of a similar length that we knowof. The setting and methods we used to maximize adherence ratespermitted a robust test of the three dietary strategies, asassessed by weight loss and established measurements of lipids,glycemic control, and hepatic and inflammatory biomarkers. Asnoted by Manzoni and colleagues, we did not assess cost-effectiveness.However, we believe that cost-effective interventions can beimplemented in the workplace, with the use of group meetingsand by working with food-service providers.
Several readers have informed us that our online SupplementaryAppendix was incomplete. We have updated the appendix to providemore complete information.
Iris Shai, R.D., Ph.D. Ben-Gurion University Beer-Sheva 84105, Israel irish{at}bgu.ac.il
Yaakov Henkin, M.D. Soroka University Medical Center Beer-Sheva 84101, Israel
Meir J. Stampfer, M.D., Dr.P.H. Harvard Medical School Boston,MA 02115
References
Shai I, Vardi H, Shahar DR, et al. Adaptation of international nutrition databases and data-entry system tools to a specific population. Public Health Nutr 2003;6:401-406. [ISI][Medline]
Rucker D, Padwal R, Li SK, Curioni C, Lau DC. Long term pharmacotherapy for obesity and overweight: updated meta-analysis. BMJ 2007;335:1194-1199. [Free Full Text]
Correction
to
Shai et al., 
Previous
