Mwc.com.br

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
Context Popular diets, particularly those low in carbohydrates, have challenged cur-
rent recommendations advising a low-fat, high-carbohydrate diet for weight loss. Po- tential benefits and risks have not been tested adequately.
Objective To compare 4 weight-loss diets representing a spectrum of low to high
carbohydrate intake for effects on weight loss and related metabolic variables.
Design, Setting, and Participants Twelve-month randomized trial conducted in
the United States from February 2003 to October 2005 among 311 free-living, over- weight/obese (body mass index, 27-40) nondiabetic, premenopausal women.
Intervention Participants were randomly assigned to follow the Atkins (n=77), Zone
(n=79), LEARN (n=79), or Ornish (n=76) diets and received weekly instruction for 2
THEONGOINGOBESITYEPI- months,thenanadditional10-monthfollow-up.
Main Outcome Measures Weight loss at 12 months was the primary outcome.
Secondary outcomes included lipid profile (low-density lipoprotein, high-density lipo- protein, and non–high-density lipoprotein cholesterol, and triglyceride levels), per- centage of body fat, waist-hip ratio, fasting insulin and glucose levels, and blood pres-sure. Outcomes were assessed at months 0, 2, 6, and 12. The Tukey studentized range tional dietary weight loss guidelines (ie, test was used to adjust for multiple testing.
energy-restricted, low in fat, high in car-bohydrate)7 have been challenged, par- Results Weight loss was greater for women in the Atkins diet group compared with
the other diet groups at 12 months, and mean 12-month weight loss was significantlydifferent between the Atkins and Zone diets (PϽ.05). Mean 12-month weight loss was as follows: Atkins, −4.7 kg (95% confidence interval [CI], −6.3 to −3.1 kg), Zone, −1.6 kg (95% CI, −2.8 to −0.4 kg), LEARN, −2.6 kg (−3.8 to −1.3 kg), and Ornish, −2.2 kg (−3.6 to −0.8 kg). Weight loss was not statistically different among the Zone, LEARN, and Ornish groups. At 12 months, secondary outcomes for the Atkins group were comparable with or more favorable than the other diet groups.
h i g h - c a r b o h y d r a t e w e i g h t - l o s s Conclusions In this study, premenopausal overweight and obese women assigned
to follow the Atkins diet, which had the lowest carbohydrate intake, lost more weight and experienced more favorable overall metabolic effects at 12 months than women assigned to follow the Zone, Ornish, or LEARN diets. While questions remain about long-term effects and mechanisms, a low-carbohydrate, high-protein, high-fat diet may be considered a feasible alternative recommendation for weight loss.
Trial Registration clinicaltrials.gov Identifier: NCT00079573
were limited by combinations of smallsample sizes, high rates of attrition, Author Affiliations: Stanford Prevention Research Cen-
ter and the Department of Medicine, Stanford Uni-
stantially different diets and 1 diet based versity Medical School, Stanford, Calif.
Corresponding Author: Christopher D. Gardner,
PhD, Hoover Pavilion, N229, 211 Quarry Rd,Stanford, CA 94305-5705 (cgardner@stanford 2007 American Medical Association. All rights reserved.
(Reprinted) JAMA, March 7, 2007—Vol 297, No. 9 969
diet books: Dr Atkins’ New Diet Revo- cific goals for energy restriction, while ships, and Nutrition; low in fat, high in lution,8 Enter the Zone,9 The LEARN Manual for Weight Management,18 or Eat were no specific energy restriction goals.
classes led by a registered dietitian once gest multiple strategies, such as relapse Participants
Participants were recruited from the local “strongly disagree” to “strongly agree,” aged 25 to 50 years were invited to enroll Process and Outcome Measures
Diet and Physical Activity Data. Di-
etary intake data were collected by tele- the other 3 diet groups. Efforts to maxi- tions); type 1 or 2 diabetes mellitus; heart, renal, or liver disease; cancer or active apolis). Data collectors were trained and month data collection, respectively.
expenditure; alcohol intake of at least 3 for 20 g/d or less of carbohydrate for “in- within the next year. Race/ethnicity data duction” (usually 2-3 months) and 50 g/d quent “ongoing weight loss” phase. The ancillary analyses of subgroups. All study drate, protein, and fat, respectively. The Anthropometric Data. Height was
Intervention
10% energy from saturated fat, caloric re- the nearest 0.1 kg on a calibrated clini- striction, increased exercise, and behav- diet book.8,9,18,19 The guidelines for the 970 JAMA, March 7, 2007—Vol 297, No. 9 (Reprinted)
2007 American Medical Association. All rights reserved.
Metabolic Measures. Blood samples
fast. Plasma total cholesterol and triglyc- established methods.22,23 High-density li- mal distributions for testing; for ease of coefficients of variation were all Յ3.1%).
Blood glucose was measured using amodification of the glucose oxidase/ Figure 1. Participant Flow Through the Trial
sured 3 times at 2-minute intervals asdescribed elsewhere30; the initial read- Statistical Analyses
ing a spectrum of carbohydrate in-take, was more effective than any other group difference in weight change was2.7 kg (6 lb, approximately 3% for a to-treat methods with baseline valuescarried forward for missing values.
2007 American Medical Association. All rights reserved.
(Reprinted) JAMA, March 7, 2007—Vol 297, No. 9 971
(PϽ.001): −497 (SD, 496), −387 (SD, were no significant interactions. All sta- groups (P = .30). Participant ratings tistical tests were 2-tailed using a sig- tober 2005. FIGURE 1 shows partici-
pant flow; TABLE 1 shows baseline
Dietary Intake and Energy
Expenditure
LEARN (P = .05) (Table 2). At subse- all) (TABLE 2). However, relative to
pattern, higher to lower intakes, wasstatistically significant for protein, fat,and saturated fat at all time points.
Table 1. Baseline Participant Characteristics*
All Diets
Characteristics
significant mean increase (PϽ.05) in energy expenditure at all time points for Weight and Anthropometric
Outcomes
−6.3 to −3.1 kg) for Atkins, −1.6 kg (95% CI, −2.8 to −0.4 kg) for Zone, −2.2 (FIGURE 2). At the 2- and 6-month in-
Abbreviations: HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
SI conversions: To convert LDL-C, HDL-C, and total cholesterol to mmol/L, multiply by 0.0259. To convert triglycerides to mmol/L, multiply by 0.0113. To convert glucose to mmol/L, multiply by 0.0555.
*Data are expressed as mean (SD) unless otherwise indicated.
†Calculated as weight in kilograms divided by height in meters squared.
groups (PϽ.05). Weight change among 972 JAMA, March 7, 2007—Vol 297, No. 9 (Reprinted)
2007 American Medical Association. All rights reserved.
Ancillary Analyses of Diet Group
Effects Independent of
mained statistically significant after in- Changes in Weight
(P = .07) or waist-hip ratio (P = .10) (TABLE 3).
using linear regression. Each of the sta- signed to follow diets having higher car- Lipid Outcomes
Results generated by 84% of the study
population (n = 262) with baseline
Table 2. Mean Dietary Intake and Energy Expenditure by Diet Group and Time Point*
P
Value†
were not significant at any time point.
Insulin, Glucose, and
Blood Pressure Outcomes
month differences were significantly dif- *Data presented are unadjusted raw data with no imputations for missing data. Standard deviations are presented in pa- rentheses. Sample sizes for baseline and 2, 6, and 12 months, respectively, are: Atkins, n = 77, 73, 71, and 68; Zone, n = 79, 73, 67, and 57; LEARN, n = 79, 73, 66, and 60; and Ornish, n = 76, 72, 67, and 56.
pressure was significantly greater for the a,b,c,d When the analysis of variance (last column) was statistically significant (PϽ.05), all pairwise comparisons among diet Atkins group than for any other group.
groups were tested for statistical significance using the Tukey studentized range test. Pairwise comparisons that weresignificantly different from one another are indicated by superscripts as follows: when the values for 2 diet groups within For diastolic pressure, the only signifi- a row do not share a common superscript, they are significantly different, whereas if the values do share a commonsuperscript, they are not significantly different.
2007 American Medical Association. All rights reserved.
(Reprinted) JAMA, March 7, 2007—Vol 297, No. 9 973
Figure 2. Weight Change Relative to
most consistent findings in recent trials observed statistically significant differ- able to at least 2 factors. One factor con- these study conditions. The triglyceride- Baseline values were carried forward for any missing values. The overall diet groupϫtime interaction wassignificant (P Ͻ.001). The analysis of variance test for differences among diet groups in weight change from baseline was significant at 2 and 6 months (PϽ.001), and at 12 months (P=.01). Analyses of all pairwisedifferences by the Tukey standardized range test (Ͻ.05) its inclusion criteria. A second likely fac- indicate that the Atkins diet group was significantly tor was differences in statistical power; different than all other diet groups at 2 and 6 monthsand that the Atkins diet group was significantly dif- ings are consistent with a beneficial in- ferent than the Zone diet group at 12 months. There were no significant differences among the Zone,LEARN, or Ornish diet groups at any time point. Er- ror bars indicate standard error of the mean.
bolic risk factors at 2 and 6 months. The finding of greater weight loss for the At- tically adjusting for weight loss differ- reaching statistical significance in com- no significant differences in weight loss level of significance was diminished.
not designed to specifically address this fat, will adversely affect blood lipid lev- in recent weight-loss diet trials. The re- cent trials, like the current study, have were greater for participants in the very- 974 JAMA, March 7, 2007—Vol 297, No. 9 (Reprinted)
2007 American Medical Association. All rights reserved.
Table 3. Mean Changes in Secondary Outcomes Relative to Baseline, by Diet Group and Time*
P Value
Overall Diet
Group ϫ Time†
Months‡
SI conversions: To convert LDL-C, HDL-C, and total cholesterol to mmol/L, multiply by 0.0259. To convert triglycerides to mmol/L, multiply by 0.0113. To convert glucose to *Intention-to-treat analysis, with baseline data carried forward for missing values. Standard deviations are presented in parentheses. For LDL-C, HDL-C, triglyceride, non−HDL-C, insulin, and glucose data, results are presented for those with available blood sample data (84% of full sample): Atkins, n = 70; Zone, n = 65; LEARN, n = 63;and Ornish, n = 64.
†P value for diet group ϫ time interaction, determined using mixed-model and autoregressive covariance structure.
‡P values for 12-month change from baseline results, determined by analysis of variance.
§Calculated as weight in kilograms divided by height in meters squared.
a,bFor a given outcome measure at the 12-month time point, when the analysis of variance (last column) was statistically significant (PϽ.05), all pairwise comparisons among diet groups were tested for statistical significance using the Tukey studentized range test. Pairwise comparisons that were significantly different from one another are indicated bysuperscripts as follows: when the values for 2 diet groups within a row do not share a common superscript, they are significantly different, whereas if the values do share acommon superscript, they are not significantly different.
2007 American Medical Association. All rights reserved.
(Reprinted) JAMA, March 7, 2007—Vol 297, No. 9 975
attrition rates, the contrast of 4 rather differed significantly in composition be- were attributable specifically to the low weight-loss trials that substituted either drate constant40,41 or protein for carbo- stant,38,42,43 the higher-protein diets led increase the external validity of the find- at least as large as for any other dietary pattern and that the lipid effects are un- regardless of macronutrient content.
Author Contributions: Drs Gardner and Balise had full
access to all of the data in the study and take respon- sibility for the integrity of the data and the accuracyof the data analysis.
Study concept and design: Gardner, Kraemer, King.
Acquisition of data: Gardner.
Analysis and interpretation of data: Gardner, Kiazand,
Alhassan, Kim, Stafford, Balise, Kraemer, King.
Drafting of the manuscript: Gardner, Kiazand, Balise,Kraemer, King.
Critical revision of the manuscript for important in- tellectual content: Gardner, Kiazand, Alhassan, Kim,Stafford, Kraemer, King.
Statistical analysis: Gardner, Alhassan, Stafford, Balise, Obtained funding: Gardner, King.
Administrative, technical, or material support: Kiazand.
Financial Disclosures: None reported.
Funding/Support: This investigation was supported by
National Institutes of Health grant R21AT1098, by a grant from the Community Foundation of Southeast-ern Michigan, and by Human Health Service grant M01-RR00070, General Clinical Research Centers, Na- tional Center for Research Resources, National Insti- level.4-6,45,46 Greater success with long- tutes of Health.
Role of the Sponsor: None of the funding agencies
played any role in the design and conduct of the study; collection, management, analysis, and interpretationof the data; or preparation, review, and approval of Acknowledgment: We gratefully acknowledge the work
of the study staff who worked with participants in re-
cruitment, intervention, and data collection, including Rise Cherin, MS, RD, Kathryn Newell, MS, Suzanne Ol-son, MS, Jennifer Morris, PhD, Jane Borchers, MS, RD, diture (ie, regular physical activity), and CONCLUSIONS
Laurie Ausserer, MS, Ellen DiNucci, MA, Kelly Boying- ton, Jana Stone, Andrea Vaccarella, RD, Noel Segali, tal factors, such as portion sizes of res- RD, and Gretchen George, MS, RD, all of Stanford Uni-versity, as well as the staff of the Stanford University Hospital General Clinical Research Center.
976 JAMA, March 7, 2007—Vol 297, No. 9 (Reprinted)
2007 American Medical Association. All rights reserved.
REFERENCES
randomized controlled trials. Arch Intern Med. 2006; loss diets. Cleve Clin J Med. 2001;68:761, 765-766, 1. Ogden CL, Carroll MD, Curtin LR, McDowell MA,
18. Brownell KD. The LEARN Manual for Weight
35. Connor WE, Connor SL. Should a low-fat, high-
Tabak CJ, Flegal KM. Prevalence of overweight and Management. Dallas, Tex: American Health Publish- carbohydrate diet be recommended for everyone? the obesity in the United States, 1999-2004. JAMA. 2006; case for a low-fat, high-carbohydrate diet. N Engl J 19. Ornish D. Eat More, Weigh Less. New York, NY:
2. Finkelstein EA, Ruhm CJ, Kosa KM. Economic causes
36. Katan MB, Grundy SM, Willett WC. Should a low-
and consequences of obesity. Annu Rev Public Health.
20. Sallis JF, Haskell WL, Wood PD, et al. Physical ac-
fat, high-carbohydrate diet be recommended for ev- tivity assessment methodology in the Five-City Project.
eryone? beyond low-fat diets. N Engl J Med. 1997;337: 3. Appel LJ, Champagne CM, Harsha DW, et al. Ef-
Am J Epidemiol. 1985;121:91-106.
fects of comprehensive lifestyle modification on blood 21. American College of Sports Medicine. Physical Fit-
37. Berneis KK, Krauss RM. Metabolic origins and clini-
pressure control: main results of the PREMIER clinical ness Testing and Interpretation. Philadelphia, Pa: Wil- cal significance of LDL heterogeneity. J Lipid Res. 2002; trial. JAMA. 2003;289:2083-2093.
4. Knowler WC, Barrett-Connor E, Fowler SE, et al.
22. Allain CC, Poon LS, Chan CS, Richmond W, Fu
38. Krauss RM, Blanche PJ, Rawlings RS, Fernstrom
Reduction in the incidence of type 2 diabetes with life- PC. Enzymatic determination of total serum cholesterol.
HS, Williams PT. Separate effects of reduced carbo- style intervention or metformin. N Engl J Med. 2002; hydrate intake and weight loss on atherogenic 23. Sampson EJ, Demers LM, Krieg AF. Faster enzy-
dyslipidemia. Am J Clin Nutr. 2006;83:1025-1031.
5. Pan XR, Li GW, Hu YH, et al. Effects of diet and
matic procedure for serum triglycerides. Clin Chem.
39. Gardner CD, Fortmann SP, Krauss RM. Associa-
exercise in preventing NIDDM in people with tion of small low-density lipoprotein particles with the impaired glucose tolerance: the Da Qing IGT and 24. Warnick GR, Albers JJ. A comprehensive evalu-
incidence of coronary artery disease in men and Diabetes Study. Diabetes Care. 1997;20:537- ation of the heparin-manganese precipitation proce- women. JAMA. 1996;276:875-881.
dure for estimating high density lipoprotein cholesterol.
40. Luscombe-Marsh ND, Noakes M, Wittert GA, Keogh
6. Stevens VJ, Obarzanek E, Cook NR, et al. Long-
JB, Foster P, Clifton PM. Carbohydrate-restricted term weight loss and changes in blood pressure: re- 25. Friedewald WT, Levy R, Fredrickson D. Estima-
diets high in either monounsaturated fat or protein sults of the Trials of Hypertension Prevention, phase tion of the concentration of low-density lipoprotein are equally effective at promoting fat loss and im- II. Ann Intern Med. 2001;134:1-11.
cholesterol in plasma, with use of the preparative proving blood lipids. Am J Clin Nutr. 2005;81:762- 7. Clinical guidelines on the identification, evalua-
ultracentrifuge. Clin Chem. 1972;18:499-502.
tion, and treatment of overweight and obesity in 26. Gardner CD, Winkleby MA, Fortmann SP. Popu-
41. Weigle DS, Breen PA, Matthys CC, et al. A high-
adults—the Evidence Report: National Institutes of lation frequency distribution of non-high-density li- protein diet induces sustained reductions in appetite, Health. Obes Res. 1998;6(suppl 2):51S-209S.
poprotein cholesterol (Third National Health and Nu- ad libitum caloric intake, and body weight despite 8. Atkins R. Dr Atkins’ New Diet Revolution. New
trition Examination Survey [NHANES III], 1988-1994).
compensatory changes in diurnal plasma leptin Am J Cardiol. 2000;86:299-304.
and ghrelin concentrations. Am J Clin Nutr. 2005;82: 9. Sears B, Lawren W. Enter the Zone. New York, NY:
27. Morgan CR, Lazarow A. Immunoassay of insu-
lin: two antibody system: plasma insulin levels in nor- 42. Farnsworth E, Luscombe ND, Noakes M, Wittert
10. Bravata DM, Sanders L, Huang J, et al. Efficacy
mal, sub diabetic, and diabetic rats. Diabetes. 1963;12: G, Argyiou E, Clifton PM. Effect of a high-protein, en- and safety of low-carbohydrate diets: a systematic ergy-restricted diet on body composition, glycemic con- review. JAMA. 2003;289:1837-1850.
28. Trinder P. Determination of blood glucose using
trol, and lipid concentrations in overweight and obese 11. Freedman MR, King J, Kennedy E. Popular
hyperinsulinemic men and women. Am J Clin Nutr.
diets: a scientific review. Obes Res. 2001;9(suppl): carcinogenic chromogen. J Clin Pathol. 1969;22:158- 43. Noakes M, Keogh JB, Foster PR, Clifton PM. Effect
12. Brehm BJ, Seeley RJ, Daniels SR, D’Alessio DA. A
29. Lott JA, Turner K. Evaluation of Trinder’s glu-
of an energy-restricted, high-protein, low-fat diet rela- randomized trial comparing a very low carbohydrate cose oxidase method for measuring glucose in serum tive to a conventional high-carbohydrate, low-fat diet diet and a calorie-restricted low fat diet on body weight and urine. Clin Chem. 1975;21:1754-1760.
on weight loss, body composition, nutritional status, and cardiovascular risk factors in healthy women. J Clin 30. King AC, Sallis JF, Dunn AL, et al. Overview of
and markers of cardiovascular health in obese women.
Endocrinol Metab. 2003;88:1617-1623.
the Activity Counseling Trial (ACT) intervention for pro- Am J Clin Nutr. 2005;81:1298-1306.
13. Foster GD, Wyatt HR, Hill JO, et al. A random-
moting physical activity in primary health care settings.
44. Appel LJ, Sacks FM, Carey VJ, et al. Effects of pro-
ized trial of a low-carbohydrate diet for obesity. N Engl Med Sci Sports Exerc. 1998;30:1086-1096.
tein, monounsaturated fat, and carbohydrate intake 31. Wood PD, Stefanick ML, Dreon DM, et al. Changes
on blood pressure and serum lipids: results of the 14. Stern L, Iqbal N, Seshadri P, et al. The effects
in plasma lipids and lipoproteins in overweight men OmniHeart randomized trial. JAMA. 2005;294:2455- of low-carbohydrate versus conventional weight during weight loss through dieting as compared with loss diets in severely obese adults: one-year fol- exercise. N Engl J Med. 1988;319:1173-1179.
45. Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity
low-up of a randomized trial. Ann Intern Med. 2004; 32. Wood PD, Stefanick ML, Williams PT, Haskell WL.
and the environment: where do we go from here? The effects on plasma lipoproteins of a prudent weight- 15. Yancy WS Jr, Olsen MK, Guyton JR, Bakst RP,
reducing diet, with or without exercise, in over- 46. Kim S, Popkin BM. Understanding the epidemi-
Westman EC. A low-carbohydrate, ketogenic diet ver- weight men and women. N Engl J Med. 1991;325:461- ology of overweight and obesity—a real global pub- sus a low-fat diet to treat obesity and hyperlipid- lic health concern. Int J Epidemiol. 2006;35:60-67.
emia: a randomized, controlled trial. Ann Intern Med.
33. National Cholesterol Education Program (NCEP)
47. Drewnowski A. Obesity and the food environ-
Expert Panel on Detection, Evaluation, and Treat- ment: dietary energy density and diet costs. Am J Prev 16. Dansinger ML, Gleason JA, Griffith JL, Selker HP,
ment of High Blood Cholesterol in Adults (Adult Treat- Schaefer EJ. Comparison of the Atkins, Ornish, Weight ment Panel III). Third Report of the National Choles- 48. Ello-Martin JA, Ledikwe JH, Rolls BJ. The influ-
Watchers, and Zone diets for weight loss and heart terol Education Program (NCEP) Expert Panel on ence of food portion size and energy density on en- disease risk reduction: a randomized trial. JAMA. 2005; Detection, Evaluation, and Treatment of High Blood ergy intake: implications for weight management. Am Cholesterol in Adults (Adult Treatment Panel III) final J Clin Nutr. 2005;82(suppl):236S-241S.
17. Nordmann AJ, Nordmann A, Briel M, et al. Ef-
report. Circulation. 2002;106:3143-3421.
49. Foster GD, Makris AP, Bailer BA. Behavioral treat-
fects of low-carbohydrate vs low-fat diets on weight 34. Blackburn GL, Phillips JC, Morreale S. Physi-
ment of obesity. Am J Clin Nutr. 2005;82(suppl):230S- loss and cardiovascular risk factors: a meta-analysis of cian’s guide to popular low-carbohydrate weight- 2007 American Medical Association. All rights reserved.
(Reprinted) JAMA, March 7, 2007—Vol 297, No. 9 977
Table 4, the Reynolds Risk Score correctly results in an ab- Financial Disclosures: Dr Ridker reports that he currently or in the past 5 years
has received research funding support from multiple not-for-profit entities includ-
solute increase in the number who would be recom- ing the National Heart, Lung, and Blood Institute, the National Cancer Institute, mended for treatment when thresholds are set at either 20% the American Heart Association, the Doris Duke Charitable Foundation, the Leducq 10-year risk or at 10% 10-year risk, thus achieving a net clini- Foundation, the Donald W. Reynolds Foundation, and the James and Polly An-nenberg La Vea Charitable Trusts. Dr Ridker also reports that currently or in the cal benefit. As with any risk classification system, perfect past 5 years he has received investigator-initiated research support from multiple prediction will not be achieved, but an overall improve- for-profit entities including AstraZeneca, Bayer, Bristol-Myers Squibb, Dade-Behring, Novartis, Pharmacia, Roche, Sanofi-Aventis, and Variagenics. Dr Ridker ment in the targeting of prescription drugs to those women reports being listed as a coinventor on patents held by the Brigham and Women’s with the most appropriate levels of risk should help maxi- Hospital that relate to the use of inflammatory biomarkers in cardiovascular dis-ease and has served as a consultant to Schering-Plough, Sanofi/Aventis, Astra- mize benefits while minimizing cost and toxicity. Wang et
Zeneca, Isis Pharmaceutical, Dade-Behring, and Vascular-Biogenics. Dr Cook re- al are also concerned about the use of self-reported blood ports having received funding from the National Heart, Lung, and Blood Institute,the National Cancer Institute, and Roche Diagnostics, and has served as a con- pressure, weight, diabetes, and smoking. However, these vari- ables show a similar magnitude of prediction in our data as 1. Hosmer DW, Lemeshow S. Goodness of fit tests for the multiple logistic re-
gression model. Commun Stat Theor Methods. 1980;A9:1043-1069.
With regard to comments from Dr Stevens and Ms Cole- 2. D’Agostino RB, Griffith JL, Schmidt CH, Terrin N. Measures for evaluating model
performance. In: American Statistical Association 1996 Proceedings of the Sec-
man, while Table 5 compares fit using the model most of- tion on Biometrics, Chicago, IL, August 1996. Alexandria, VA: American Statisti- ten used in clinical practice, Table 4 shows superiority of cal Association; 1997:253-258.
3. Salomaa V, Harold K, Sundvill J, Jousilahti P. Brain natriuretic peptide as a pre-
the new models built using the same population and out- dictor of coronary and cardiovascular events and all-cause deaths in general popu- come definition. We acknowledge that external validation, lation [abstract]. Circulation. 2007;115(8):e269.
using different cohorts, would be a useful next step. It istrue that the Hosmer-Lemeshow statistic can be consid-ered a general measure of goodness of fit.1 However, since CORRECTIONS
it directly compares observed with expected events, it is more Incorrect Wording and Data Error: In the Original Contribution entitled “Com-
sensitive to recalibration than most other measures, par- parison of the Atkins, Zone, Ornish, and LEARN Diets for Change in Weight and ticularly the c-statistic, and is often treated as a measure of Related Risk Factors Among Overweight Premenopausal Women: The A TO ZWeight Loss Study: A Randomized Trial” published in the March 7, 2007, issue of JAMA (2007;297(9):969-977), a sentence was incorrectly worded in the ab- We do not concur with Dr Daniels and colleagues that stract, and data were reported incorrectly in the text. On page 969, in the “Con-clusions” section of the abstract, the first sentence should have read “In this study, epidemiologic data on natriuretic peptides support the premenopausal overweight and obese women assigned to follow the Atkins diet, use of this biomarker in healthy populations. Of the which had the lowest carbohydrate intake, had lost more weight at 12 monthsthan those assigned to the Zone diet, and had experienced comparable or more articles cited, most included prevalent myocardial infarc- favorable metabolic effects than those assigned to follow the Zone, Ornish, or LEARN tion at baseline or evaluated elderly cohorts without diets.” On page 972, in the last paragraph, the mean 12-month weight changesfor the LEARN and Ornish diets were reversed: for LEARN it should have been adequate exclusion of prior cardiovascular events. More −2.6 kg (95% CI, −3.8 to −1.3 kg) and for Ornish it should have been −2.2 kg recent data suggest that B-type natriuretic peptide does not predict cardiovascular events among those free of dis- Incorrect Prevalence: In the Editorial entitled “Mandatory HPV Vaccination: Pub-
lic Health vs Private Wealth” published in the May 2, 2007, issue of JAMA (2007;297(17):1921-1923), 2 sentences regarding HPV prevalence were inaccurate. On Paul M Ridker, MD
page 1921, in the second paragraph, the second to last sentence should read: “Al- [email protected]
though infection with high-risk HPV types . . . high-risk types 16 and 18 have a Nancy R. Cook, ScD
relatively low prevalence (2.3% among screened females),4 and not all wom-en. . . . ” Also on page 1921, second column, the last paragraph on the page should Brigham and Women’s Hospital
read: “Given that the overall prevalence of HPV vaccine types associated with cer- Boston, Massachusetts
vical cancer is relatively low (2.3%). . . . ” 178 JAMA, July 11, 2007—Vol 298, No. 2 (Reprinted)
2007 American Medical Association. All rights reserved.
Table 4, the Reynolds Risk Score correctly results in an ab- Financial Disclosures: Dr Ridker reports that he currently or in the past 5 years
has received research funding support from multiple not-for-profit entities includ-
solute increase in the number who would be recom- ing the National Heart, Lung, and Blood Institute, the National Cancer Institute, mended for treatment when thresholds are set at either 20% the American Heart Association, the Doris Duke Charitable Foundation, the Leducq 10-year risk or at 10% 10-year risk, thus achieving a net clini- Foundation, the Donald W. Reynolds Foundation, and the James and Polly An-nenberg La Vea Charitable Trusts. Dr Ridker also reports that currently or in the cal benefit. As with any risk classification system, perfect past 5 years he has received investigator-initiated research support from multiple prediction will not be achieved, but an overall improve- for-profit entities including AstraZeneca, Bayer, Bristol-Myers Squibb, Dade-Behring, Novartis, Pharmacia, Roche, Sanofi-Aventis, and Variagenics. Dr Ridker ment in the targeting of prescription drugs to those women reports being listed as a coinventor on patents held by the Brigham and Women’s with the most appropriate levels of risk should help maxi- Hospital that relate to the use of inflammatory biomarkers in cardiovascular dis-ease and has served as a consultant to Schering-Plough, Sanofi/Aventis, Astra- mize benefits while minimizing cost and toxicity. Wang et
Zeneca, Isis Pharmaceutical, Dade-Behring, and Vascular-Biogenics. Dr Cook re- al are also concerned about the use of self-reported blood ports having received funding from the National Heart, Lung, and Blood Institute,the National Cancer Institute, and Roche Diagnostics, and has served as a con- pressure, weight, diabetes, and smoking. However, these vari- ables show a similar magnitude of prediction in our data as 1. Hosmer DW, Lemeshow S. Goodness of fit tests for the multiple logistic re-
gression model. Commun Stat Theor Methods. 1980;A9:1043-1069.
With regard to comments from Dr Stevens and Ms Cole- 2. D’Agostino RB, Griffith JL, Schmidt CH, Terrin N. Measures for evaluating model
performance. In: American Statistical Association 1996 Proceedings of the Sec-
man, while Table 5 compares fit using the model most of- tion on Biometrics, Chicago, IL, August 1996. Alexandria, VA: American Statisti- ten used in clinical practice, Table 4 shows superiority of cal Association; 1997:253-258.
3. Salomaa V, Harold K, Sundvill J, Jousilahti P. Brain natriuretic peptide as a pre-
the new models built using the same population and out- dictor of coronary and cardiovascular events and all-cause deaths in general popu- come definition. We acknowledge that external validation, lation [abstract]. Circulation. 2007;115(8):e269.
using different cohorts, would be a useful next step. It istrue that the Hosmer-Lemeshow statistic can be consid-ered a general measure of goodness of fit.1 However, since CORRECTIONS
it directly compares observed with expected events, it is more Incorrect Wording and Data Error: In the Original Contribution entitled “Com-
sensitive to recalibration than most other measures, par- parison of the Atkins, Zone, Ornish, and LEARN Diets for Change in Weight and ticularly the c-statistic, and is often treated as a measure of Related Risk Factors Among Overweight Premenopausal Women: The A TO ZWeight Loss Study: A Randomized Trial” published in the March 7, 2007, issue of JAMA (2007;297(9):969-977), a sentence was incorrectly worded in the ab- We do not concur with Dr Daniels and colleagues that stract, and data were reported incorrectly in the text. On page 969, in the “Con-clusions” section of the abstract, the first sentence should have read “In this study, epidemiologic data on natriuretic peptides support the premenopausal overweight and obese women assigned to follow the Atkins diet, use of this biomarker in healthy populations. Of the which had the lowest carbohydrate intake, had lost more weight at 12 monthsthan those assigned to the Zone diet, and had experienced comparable or more articles cited, most included prevalent myocardial infarc- favorable metabolic effects than those assigned to follow the Zone, Ornish, or LEARN tion at baseline or evaluated elderly cohorts without diets.” On page 972, in the last paragraph, the mean 12-month weight changesfor the LEARN and Ornish diets were reversed: for LEARN it should have been adequate exclusion of prior cardiovascular events. More −2.6 kg (95% CI, −3.8 to −1.3 kg) and for Ornish it should have been −2.2 kg recent data suggest that B-type natriuretic peptide does not predict cardiovascular events among those free of dis- Incorrect Prevalence: In the Editorial entitled “Mandatory HPV Vaccination: Pub-
lic Health vs Private Wealth” published in the May 2, 2007, issue of JAMA (2007;297(17):1921-1923), 2 sentences regarding HPV prevalence were inaccurate. On Paul M Ridker, MD
page 1921, in the second paragraph, the second to last sentence should read: “Al- [email protected]
though infection with high-risk HPV types . . . high-risk types 16 and 18 have a Nancy R. Cook, ScD
relatively low prevalence (2.3% among screened females),4 and not all wom-en. . . . ” Also on page 1921, second column, the last paragraph on the page should Brigham and Women’s Hospital
read: “Given that the overall prevalence of HPV vaccine types associated with cer- Boston, Massachusetts
vical cancer is relatively low (2.3%). . . . ” 178 JAMA, July 11, 2007—Vol 298, No. 2 (Reprinted)
2007 American Medical Association. All rights reserved.

Source: http://www.mwc.com.br/files/Gardner_-_Standford_A_to_Z.pdf