sign in sign up
<<
Home , Weight loss

SECTION III

Is Weight Loss Beneficial for Reduction of Morbidity and Mortality? What is the controversy about?

PETER M. NILSSON, MD POSSIBLE EXPLANATIONS FOR INCREASED RISKS ASSOCIATED WITH The increase of and type 2 on a global scale has increased the interest in how to WEIGHT LOSS — As weight loss has counteract this epidemic. Improved lifestyle in general is a fundamental approach, but other been associated not only with health ben- remedies such as specific weight reduction or diabetes preventive drugs and have also efits, but also with health hazards (Table been tested. One problem to understand is what really happens after weight loss. Ongoing 1), it is important to find some putative studies will try to address this question, such as the Swedish Obese Subjects (SOS) surgery study, explanations for these paradoxical effects. the Look AHEAD (Action for Health in Diabetes) trial in the U.S. (recruiting obese type 2 diabetic One explanation is simple; the associa- patients), and the Comprehensive Rimonabant Evaluation Study of Cardiovascular End Points tions could be the result of residual con- and Outcomes (CRESCENDO) trial (by use of rimonabant versus placebo). This is very impor- tant, since previously, several observational studies in large population-based cohorts have founding and therefore be spurious and indicated some detrimental effects of weight loss, even after intentional weight loss, with in- classified as artifacts. It is well known that creased morbidity and mortality rates. chronic devastating will lead to weight loss in many patients, for example, Diabetes Care 31 (Suppl. 2):S278–S283, 2008 cardiac or advanced chronic ob- structive pulmonary disease, and thus as- besity is a well-established risk fac- confounding, 2) subclinical disease or de- sociate observational weight loss with tor for many chronic disorders such pression causing both weight loss and in- increased mortality risk. In the studies O as (CVD), creased mortality risk, 3) true detrimental from the Nordic countries (7–10), re- , and certain , as effects on by weight loss per searchers have however tried to avoid this well as for increased all-cause mortality se, and 4) weight loss being a marker of fallacy by excluding unhealthy subjects at risk (1–3). It is therefore paradoxical that early general aging (biological involution) baseline for follow-up analyses, as well as weight loss, whether only observational in susceptible individuals, causing pre- excluding the first few years of follow-up or even intentional (in some observational mature morbidity and mortality. It seems regarding mortal events to avoid con- studies) is also associated with an in- that weight loss may well improve medi- founding by subclinical disease. Another explanation is that psychiatric conditions creased mortality risk (4–10). This has cal symptoms, risk factor levels, and qual- such as depression, eventually leading to now been documented from four Nordic ity of life for obese subjects, but the countries (7–10) in recent epidemiologi- suicide, could influence the association paradoxical finding of an increased mor- cal studies based on population-based co- between weight loss due to poor tality risk associated with observational horts followed over time, after an initial in depressed subjects with later mortality assessment of weight loss compared with and even intentional weight loss (10) calls risk. This could well be true for some se- weight stability or weight increase during for more research efforts and a more cau- lect subjects, but cannot be accepted as a the first few years of follow-up. Therefore, tious attitude toward the healthy but more general explanation for larger a clinical controversy exists. obese subject asking for medical advice. groups of people experiencing weight In Malmo¨, Sweden, several studies An alternative approach is to promote loss. A third explanation is based on the have shown the health hazards related to weight stabilization as a goal for the fact that biological involution (e.g., re- obesity (11,12), but also have confirmed healthy but or even mildly duced weight and height) in healthy sub- the association between observational obese individual, because a stable weight jects is part of a normal aging process. weight loss and increased long-term mor- has been associated with less cardiovascu- This is supposed to take a more rapid tality risk in middle-aged men (7). Why is lar risk and less mortality in observational course in subjects showing signs of early there such a paradoxical association? Pos- studies, both in comparison with weight ageing, thereby increasing the risk of ear- sible explanations could include the fol- increase and weight loss (7–10). ly-onset mortality. For CVD, this can be lowing: 1) an artifact due to residual called the early vascular aging syndrome. ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● Finally, it is difficult to rule out the From the Department of Clinical Sciences, University Hospital, Malmo¨, Sweden. possibility that weight loss per se could be Address correspondence and reprint requests to Peter M. Nilsson, Department of Clinical Sciences, hazardous to health, at least in some sus- University Hospital, S-205 02 Malmo¨, Sweden. E-mail: [email protected]. The author of this article has no relevant duality of interest to declare. ceptible individuals. One piece of evi- This article is based on a presentation at the 1st World Congress of Controversies in Diabetes, Obesity and dence supporting this hypothesis is the (CODHy). The Congress and the publication of this article were made possible by unrestricted well-known risk of cholelithiasis attacks educational grants from MSD, Roche, sanofi-aventis, Novo Nordisk, Medtronic, LifeScan, World Wide, Eli in obese patients after rapid intentional Lilly, Keryx, Abbott, Novartis, Pfizer, Generx Biotechnology, Schering, and Johnson & Johnson. Abbreviations: CVD, cardiovascular disease; SOS, Swedish Obese Subjects. weight loss (13). If correct, this hypothe- DOI: 10.2337/dc08-s268 sis based on observation could be extrap- © 2008 by the American Diabetes Association. olated to other unwanted health risks and

S278 DIABETES CARE, VOLUME 31, SUPPLEMENT 2, FEBRUARY 2008 Nilsson eventually an increased mortality risk in a creased risk of both micro- and macro- LIFESTYLE INTERVENTION subset of people, even after intentional vascular complications (17,20). Similar TO DECREASE RISK OF weight loss. To settle this research di- findings have been noted for the weight DIABETES — In the Finnish Diabetes lemma based on observations, it takes increase, but at the same time, de- Prevention Study, 522 subjects with im- well-designed, randomized, controlled creased risk after smoking cessation or paired glucose tolerance participated intervention studies in large groups of by use of ␤-receptor blocking agents af- (23). The intervention group showed sig- overweight/obese subjects who are losing ter myocardial infarction (secondary nificantly greater improvement in each weight, either after drug treatment or sur- prevention) has been noted. In both intervention goal of weight loss and phys- gery, or a combination of these modes of conditions, the cardiovascular risk is ical . After 1 and 3 years, weight interventions. more or less decreased, forming a para- reductions were 4.5 and 3.5 kg in the in- dox of –associated clinical tervention group and 1.0 and 0.9 kg in RISK OF CVD AND benefits and risk reduction. the control group, respectively. Measures DIABETES IN RELATION TO of glycemia improved more in the inter- OBESITY — Overweight and obesity vention group. The intensive lifestyle in- are major contributors to both type 2 WEIGHT CONTROL tervention produced long-term beneficial diabetes and CVD. Moreover, individu- ACHIEVEMENT IN TYPE 2 changes in , physical activity, and als with type 2 diabetes who also are DIABETES — The common practice clinical and biochemical parameters. It obese are at particularly high risk for to advise diabetes patients to lose weight also substantially reduced the risk of dia- CVD morbidity and mortality, since is based on beliefs of potential benefits betes by 58%, an effect that was at least other risk factors such as hypertension documented in observational studies only partly preserved during long-term post- and dyslipidemia tend to cluster with (21,22). In addition, it may also be diffi- trial follow-up, as recently described (24). obesity (14). Several studies have This is a proof of concept that an effective cult to combine this goal of weight loss shown that obesity gradually increases intervention, including weight loss, could with that of improved glycemic control if the risk of type 2 diabetes, especially if improve glucose metabolism. However, it weight-promoting antidiabetes drugs are located in the abdominal region is difficult to disentangle beneficial effects (15,16). This risk depends on several used (insulin, sulfonylureas, or glita- associated with weight loss from effects pathophysiological mechanisms in- zones). Regretfully enough, we presently caused by other simultaneous interven- cluding decreased insulin sensitivity lack data on clinical end points from ran- tions, especially that of increased physical and secondary impairment of ␤-cell domized intervention trials to really sup- activity. function associated with obesity, the port such common advice to lose weight latter often being a consequence of the in these patients. There is no doubt that DRUG INTERVENTION TO former when (glucotox- weight control and weight reduction can DECREASE THE RISK OF icity) gradually increases and impairs reduce the risk of developing diabetes in DIABETES: ONE EXAMPLE —A the function of the ␤-cell. It is less subjects with impaired glucose tolerance, reduction in the incidence of type 2 dia- known that diabetes treatment itself as shown both by lifestyle interventions betes with lifestyle changes has previously may also increase the degree of obesity. (23–25) and by use of drugs such as orl- been demonstrated in several studies, The reason for this is the weight- istat (26,27), acarbose (28), and rosiglita- based on, for example, physical exercise. increasing effects of treatment by vari- zone (29). However, the feasibility and Addition of a weight-reducing drug to ous antidiabetes drugs, for example, by benefits by weight reduction in estab- lifestyle changes may lead to an even insulin and sulfonylureas, as shown in lished type 2 diabetes is less well docu- greater decrease in body weight and thus the U.K. Prospective Diabetes Study mented and also sometimes hard to the incidence of type 2 diabetes in obese (17), or by glitazone treatment (18,19). achieve if most antidiabetes drugs act by patients. In a 4-year double-blind pro- Paradoxically, the increased weight in increasing weight, with metformin being spective study, 3,305 patients were ran- these patients is associated with a de- the only exception. domized to lifestyle changes plus either

Table 1—The pros and cons of weight loss based on observational and intervention studies Health benefits of weight loss Lower levels of cardiovascular risk factors, e.g., blood pressure, lipids, and glycemia Less abdominal obesity increases adiponectin levels and insulin sensitivity Lower incidence of new-onset type 2 diabetes in subjects with impaired glucose tolerance or impaired glucose when combined with physical activity Reduced symptoms of musculoskeletal and joint pain, obstipation, and psychological complaints Improved quality of life in very obese subjects Health hazards of weight loss Increased mortality risk in observational studies, even in subjects with the intention to lose weight. Controversies still exist based on lack of information, e.g., on the intentionality of weight loss in other studies with opposite findings Increased risk of osteoporotic fractures in lean elderly subjects Increased risk of gallstone attacks if (intentional) weight loss is rapid Loss of muscle tissue if weight loss is only based on and not the combination with increased physical activity

DIABETES CARE, VOLUME 31, SUPPLEMENT 2, FEBRUARY 2008 S279 Is weight loss beneficial for reduction of morbidity?

120 mg or placebo, three times heart disease equivalent or two major risk care reimbursement for diabetes educa- daily (27). Participants had a BMI Ն30 factors for CVD. tion related to nutrition and physical kg/m2 and normal (79%) or impaired activity) interventions, with usual care as (21%) glucose tolerance. Primary end OBSERVATIONAL STUDY ON a control group (36). Modest weight loss points were time to onset of type 2 diabe- THE EFFECTS OF WEIGHT occurred after 6 months among intensive- tes and change in body weight. Of orl- LOSS IN DIABETES — If random- lifestyle participants and was greater than istat-treated patients, 52% completed ized controlled intervention studies are the weight loss among usual-care partici- treatment compared with 34% of placebo not available, the second best option on pants (2.6 vs. 0.4 kg, P Ͻ 0.01). At 12 recipients (P Ͻ 0.0001). After 4 years’ which to base decisions in clinical medi- months, a greater proportion of intensive- treatment, the cumulative incidence of di- cine are well-designed and large observa- lifestyle participants had lost Ն2 kg than abetes was 9.0% with placebo and 6.2% tional studies. In one such study, medical the usual-care participants (49 vs. 25%, with orlistat therapy (“intention to treat”), records were reviewed of all 263 patients P Ͻ 0.05). No differences in weight corresponding to a risk reduction of 37% with type 2 diabetes from a diabetes clinic change were observed between reimburs- (P ϭ 0.0032). The preventive effect was in Scotland who were known to have died able-lifestyle and usual-care participants. explained by the difference in subjects in the mid-1980s (22). Mean age was 65 Glycated hemoglobin was reduced with impaired glucose tolerance. Mean years (range 57–75) at diagnosis and 72 among all groups (P Ͻ 0.05) in the same weight loss after 4 years was significantly years (range 66–80) for men and 75 years way. In conclusion, this study showed greater with orlistat (5.8 vs. 3.0 kg with (range 72–83) for women, at death. Anal- that improvement in both weight and placebo; P Ͻ 0.001) and similar between ysis of survival in 233 patients who lived glycemia are indeed attainable by life- orlistat recipients with impaired (5.7 kg) Ͼ1 year (189 overweight), using stepwise style interventions designed for individ- or normal glucose tolerance (5.8 kg) at multiple regression analysis, indicated uals with type 2 diabetes living in rural baseline. Compared with lifestyle changes the following as significant (P Ͻ 0.05) ad- communities, often in less favorable so- alone, orlistat plus lifestyle changes thus verse independent variables: age at diag- cial conditions. resulted in a greater reduction in the inci- nosis, presence of clinical ischemic heart However, we need data from larger dence of type 2 diabetes over 4 years and disease at diagnosis, and plasma glucose studies on the methodology and clinical produced greater weight loss in this obese at diagnosis. Significant favorable vari- outcomes of intentional weight loss in at- population. However, a difference in dia- ables were oral hypoglycemic drug ther- risk individuals with diabetes. Fortu- betes incidence was detectable only in the apy, weight loss in the first year, and an nately, a few clinical trials are now impaired glucose tolerance subgroup. interaction between weight loss and BMI ongoing and will hopefully be able to in- for patients with BMI Ͼ25 kg/m2. crease the knowledge on how to achieve Changes in treatment over the years are weight loss in obese subjects and diabetic EFFECTS OF RIMONABANT likely to have biased these results toward patients, as well as the long-term conse- ON WEIGHT LOSS AND including oral hypoglycemic therapy and quences of this achievement for clinical CLINICAL EFFECTS — A new alter- excluding the expected adverse effect of end points. native to lose weight and to improve risk smoking. Mean weight loss at 1 year was factor control is the endocannaboid-1 re- 2.6 kg for individuals with BMI 25–30 LOOK AHEAD: ceptor antagonist rimonabant. Clinical kg/m2 and 6.8 kg for individuals with BMI INTERVENTION FOR studies in abdominally obese subjects Ͼ30 kg/m2, following dietary advice. For WEIGHT LOSS IN TYPE 2 have documented weight loss, improved the average patient, each 1-kg weight loss DIABETES — Although short-term glucose metabolism, and lipid control, as was associated with 3–4 months’ pro- weight loss has been shown to ameliorate well as reduced blood pressure in patients longed survival. Some rest confounding obesity-related metabolic abnormalities with type 2 diabetes (30–33). Some ad- could, however, be present as this was not and CVD risk factors, the long-term con- verse effects on mental function have a randomized, controlled study. sequences of intentional weight loss in been noticed in some patients, and that is obese individuals with type 2 diabetes why this drug should not be prescribed to ROLE OF LIFESTYLE AND have previously not been adequately ex- patients with a medical history of depres- EMPOWERMENT OF amined. The primary objective of the on- sion or pronounced mental symptoms. DIABETIC PATIENTS — The ef- going Look AHEAD (Action for Health in Other effects seen in some but not all forts to help patients with established Diabetes) clinical trial (37) is to assess the studies include increased rates of smok- type 2 diabetes to improve their lifestyle long-term effects (up to 11.5 years) of an ing cessation. It is important that rimon- must often be combined with methods to intensive weight loss program delivered abant is currently being evaluated for increase self-efficacy and the empower- over 4 years in overweight and obese effects on cardiovascular morbidity and ment of the patient (35). There could be individuals with type 2 diabetes. Ap- mortality end points versus placebo in a obstacles to this ambition, such as poverty proximately 5,000 male and female par- randomized controlled study, the Com- or poor education on behalf of the patient, ticipants (45–74 years old) who have prehensive Rimonabant Evaluation Study or a lack of social support. One project type 2 diabetes and a BMI Ն25 kg/m2 of Cardiovascular End Points and Out- evaluated lifestyle interventions for pa- have been randomized to one of the two comes (CRESCENDO) study, with ex- tients with diabetes living in poor rural groups. The study is now fully recruited pected results in 2011 (34). This trial is communities. It was a 12-month random- and ongoing. The intensive lifestyle in- recruiting patients Ն55 years, with inclu- ized clinical trial (n ϭ 152) of “intensive- tervention is designed to achieve and sion criteria: waist circumference Ͼ102 lifestyle” and “reimbursable-lifestyle” maintain weight loss through decreased cm (40 inches) in males, Ͼ88 cm (35 (intensive-lifestyle intervention delivered caloric intake and increased physical inches) in females, with one coronary in the limited time associated with Medi- activity, by support of trained instruc-

S280 DIABETES CARE, VOLUME 31, SUPPLEMENT 2, FEBRUARY 2008 Nilsson tors. This program is compared to a the surgical group, while the incidence of use of ␤-receptor blockers for secondary control condition (standard care) of di- hypertension, remarkably enough, was prevention after myocardial infarction abetes support and education. The pri- equal in the two treatment groups (OR (41). Even if a weight increase is likely to mary study outcome variable is time to 1.01, 95% CI 0.61–1.67). The reason for occur after the use of these drugs, thus incidence of a major CVD event. The this is not clear but could because weight providing harder conditions for any study is designed to provide a 0.90 loss in grossly obese people in adult life weight loss, the cardiovascular risk is de- probability of detecting an 18% differ- may not be able to fully abolish the patho- creased and that is what matters the most. ence in major CVD event rates between physiological mechanisms regulating Randomized controlled trials such as the two groups. Other outcomes in- blood pressure control, set at a much Look AHEAD (37) and CRESCENDO clude components of CVD risk, cost and younger age, possibly during childhood. (34) will hopefully contribute to our un- cost-effectiveness, diabetes control and After 8 years, the maintained weight loss derstanding of the effects of intentional complications, hospitalizations, inter- was still 20.1 Ϯ 15.7 kg (16.3 Ϯ 12.3%). weight loss and thereby resolve the con- vention processes, and quality of life. In Thus, a differentiated risk factor response troversy that exists. Surgical interventions summary, the Look AHEAD trial is an was identified based on the fact that a cannot and will never be a remedy for the important clinical trial and a step to- maintained weight reduction of 16% vast majority of overweight/obese sub- ward increased evidence-based knowl- strongly counteracted the development of jects, who have increased risks for diabe- edge on how to handle obesity diabetes over 8 years but showed no long- tes, CVD, and early mortality. Still, problems in patients with established term effect on the incidence of hyperten- surgery is useful for subjects from the far type 2 diabetes. sion. Many patients with diabetes taking end of the obese distribution of BMI, pro- part in the study have also been able to viding them with improved quality of life, IS WEIGHT-REDUCING reduce or completely stop their medica- symptom relief, and a better prognosis, SURGERY THE SOLUTION? — tion (38,39) because of improved glyce- according to findings in the SOS study Another approach for weight reduction is mic control after surgery. In 2007, the (40). In the end, we need several options to use surgical intervention, such as gas- main SOS study was published (40). It for weight loss strategies, based on evi- tric bypass or gastric banding. The Swed- showed that following in dence, in patients in need of these inter- ish Obese Subjects (SOS) study is obese subjects with BMI Ͼ35 kg/m2,to- ventions. However, a strategy of weight conducted to determine whether obese tal mortality and myocardial infarction, stabilization, with conversation of muscle patients actually can improve their mor- and also events, were significantly tissue that is sometimes endangered by tality risk by losing weight (38,39). After reduced but surprisingly not stroke inci- weight loss and weight cycling, might still an initial screening study, a total of 6,328 dence. The authors concluded that bari- be a reasonable and achievable goal for subjects were recruited and extensively atric surgery for severe obesity is many overweight or even mildly obese characterized for background factors and associated with long-term weight loss and subjects who are otherwise healthy. other medical treatments. In the following decreased overall mortality. Surgical in- intervention study, 2,010 of the subjects tervention is, however, less well proven in underwent surgery for obesity (gastric other categories of overweight or in pa- References banding, gastroplasty, gastric bypass), tients not as obese as those recruited in 1. Stamler J, Vaccaro O, Neaton JD, Went- while 2,037 chose a conventional form of the SOS study. worth D: Diabetes, other risk factors, and 12-yr cardiovascular mortality for men treatment and acted as control subjects screened in the Multiple Risk Factor In- (matched pairs) during a period of up to CONCLUSIONS AND tervention Trial. Diabetes Care 16:434– 20 years. Thus, the study was nonran- IMPLICATIONS — A controversy 444, 1993 domized but well controlled, even for still exists regarding the role of intentional 2. Yusuf S, Hawken S, Ounpuu S, Bautista L, personality and psychological factors. Af- weight loss for prognosis in nondiabetic Franzosi MG, Commerford P, Lang CC, ter 10 years, the control subjects had healthy subjects, as well as in patients Rumboldt Z, Onen CL, Lisheng L Tanom- gained an average of 1.4 kg in weight. The with type 2 diabetes, when weight loss is sup S, Wangai P, Razak F, Sharma AM, surgical subjects, in contrast, showed a often a contradiction to the weight in- Anand SS, INTERHEART Study Investi- substantial and persisting decrease in crease followed by use of some anti- gators: Obesity and the risk of myocardial weight, averaging 20–30 kg, differing diabetic drugs. Without evidence from infarction in 27,000 participants from 52 countries: a case-control study. Lancet with the surgical methods used. In com- trials, we will never be able to make jus- 366:1640–1649, 2005 parison with the control group, this sur- tified clinical decisions on how and when 3. Bray GA, Bellanger T: Epidemiology, gical intervention group showed a clear weight loss should be recommended to trends, and morbidities of obesity and the decrease in the incidence of cardiovascu- risk patients and which methods should . Endocrine 29:109– lar risk factors (among others, hyperten- be preferred—lifestyle, drugs, or surgical 117, 2006 sion, hypertriglyceridemia, and diabetes) intervention. Observational studies only 4. Andres R, Muller DC, Sorkin JD: Long- as well as an improvement in cardiac represent one first step and are not fully term effects of change in body weight on function parameters and health-related reliable, since rest confounding may all-cause mortality: a review. Ann Intern quality of life. Compared with weight sta- sometimes bias the outcome. For exam- Med 119:737–743, 1993 bility, large intentional weight loss thus ple, the background drug medication 5. Pamuk ER, Williamson DF, Serdula MK, Madans J, Byers TE: Weight loss and sub- resulted in substantial reductions in the given to patients losing weight may either sequent death in a cohort of U.S. adults. 2-year incidence of several cardiovascular increase or decrease the likelihood of Ann Intern Med 119:744–748, 1993 risk factors. After 8 years, there was still a achieving weight loss, but at the same 6. Corrada MM, Kawas CH, Mozaffar F, Pa- reduced risk of developing diabetes (odds time reduce risk of morbidity and mortal- ganini-Hill A: Association of body mass ratio [OR] 0.16, 95% CI 0.07–0.36) in ity. One well-known example of this is the index and weight change with all-cause

DIABETES CARE, VOLUME 31, SUPPLEMENT 2, FEBRUARY 2008 S281 Is weight loss beneficial for reduction of morbidity?

mortality in the elderly. Am J Epidemiol 20. Dormandy JA, Charbonnel B, Eckland tion) Trial Investigators, Gerstein HC, 163:938–949, 2006 DJA, Erdmann E, Massi-Benedetti M, Yusuf S, Bosch J, Pogue J, Sheridan P, 7. Mikkelsen KL, Heitmann BL, Keiding N, Moules IK, Skene AM, Tan MH, Lefebvre Dinccag N, Hanefeld M, Hoogwerf B, Sorensen TI: Independent effects of sta- PJ, Murray GD, Standl E, Wilcox RG, Wil- Laakso M, Mohan V, Shaw J, Zinman B, ble and changing body weight on total helmsen L, Betteridge J, Birkeland K, Go- Holman RR: Effect of rosiglitazone on the mortality. Epidemiology 10:671–678, lay A, Heine RJ, Koranyi L, Laakso M, frequency of diabetes in patients with im- 1999 Mokan M, Norkus A, Pirags V, Podar T, paired glucose tolerance or impaired fast- 8. Nilsson PM, Nilsson J-Å, Hedblad B, Ber- Scheen A, Scherbaum W, Schernthaner ing glucose: a randomised controlled trial. glund G, Lindga¨rde F: The enigma of in- G, Schmitz O, Skrha J, Smith U, Taton J, Lancet 368:1096–1105, 2006 creased mortality associated with weight- PROactive Investigators: Secondary pre- 30. Van Gaal LF, Rissanen AM, Scheen AJ, loss in middle-aged men: the Malmo¨ vention of macrovascular events in pa- Ziegler O, Rossner S, RIO-Europe Study Prevention Project. J Intern Med 252:70– tients with type 2 diabetes in the Group: Effects of the cannabinoid-1 re- 78, 2002 PROactive Study (PROspective pioglitA- ceptor blocker rimonabant on weight re- 9. Dro¨yvold W, Lydersen S, Midthjell K, Nil- zone Clinical Trial In macroVascular duction and cardiovascular risk factors in sen TIL, Nilsson PM, Nilsson J-Å, Holmen Events): a randomised controlled trial. overweight patients: 1-year experience J: Weight change and mortality: the Nord- Lancet 366:1279–1289, 2005 from the RIO-Europe study. Lancet 365: Trøndelag Health Study. J Intern Med 257: 21. Mayer-Davis EJ, Costacou T: Obesity and 1389–1397, 2005 338–345, 2005 : modifiable risk factors 31. Despres JP, Golay A, Sjo¨stro¨m L: Rimon- 10. Sorensen TI, Rissanen A, Korkeila M, for prevention of type 2 diabetes. Curr abant in Obesity-Lipids Study Group: Ef- Kaprio J: Intention to lose weight, weight Diab Rep 1:170–176, 2001 fects of rimonabant on metabolic risk changes, and 18-y mortality in over- 22. Lean ME, Powrie JK, Anderson AS, Garth- factors in overweight patients with dyslip- weight individuals without co-morbidi- waite PH: Obesity, weight loss and prog- idemia. N Engl J Med 353:2121–2134, ties. PLoS Med 2:e171, 2005 nosis in type 2 diabetes. Diabet Med 2005 11. Jonsson S, Hedblad B, Engstro¨m G, Nil- 7:228–233, 1990 32. Pi-Sunyer FX, Aronne LJ, Heshmati HM, sson P, Berglund G, Janzon L: Influence 23. Lindstro¨m J, Louheranta A, Mannelin M, Devin J, Rosenstock J, RIO-North Amer- of obesity on cardiovascular risk: twen- Rastas M, Salminen V, Eriksson J, Uus- ica Study Group: Effect of rimonabant, a ty-three year follow-up of 22 025 men itupa M, Tuomilehto J, Finnish Diabetes cannabinoid-1 receptor blocker, on from an urban Swedish population. Int J Prevention Study Group: The Finnish Di- weight and cardiometabolic risk factors in Obes Relat Metab Disord 26:1046–1053, abetes Prevention Study (DPS): lifestyle overweight or obese patients: RIO-North 2002 intervention and 3-year results on diet America: a randomized controlled trial. 12. Hedblad B, Jonsson S, Nilsson P, Eng- and physical activity. Diabetes Care JAMA 295:761–775, 2006 stro¨m G, Berglund G, Janzon L: Obesity 26:3230–3236, 2003 33. Scheen AJ, Finer N, Hollander P, Jensen and myocardial infarction: vulnerability 24. Knowler WC, Barrett-Connor E, Fowler MD, Van Gaal LF, RIO-Diabetes Study related occupational level and marital sta- SE, Hamman RF, Lachin JM, Walker EA, Group: Efficacy and tolerability of rimon- tus: a 23-year follow-up of an urban Nathan DM: Diabetes Prevention Pro- abant in overweight or obese patients Swedish population. J Intern Med 252: gram Research Group: Reduction in the with type 2 diabetes: a randomized con- 542–550, 2002 incidence of type 2 diabetes with lifestyle trolled study. Lancet 368:1660–1672, 13. Tsai CJ, Leitzmann MF, Willett WC, Gio- intervention or metformin. N Engl J Med 2006 vannucci EL: Weight cycling and risk of 346:393–403, 2002 34. CRESCENDO Comprehensive Rimon- gallstone disease in men. Arch Intern Med 25. Lindstrom J, Ilanne-Parikka P, Peltonen abant Evaluation Study of Cardiovascular 166:2369–2374, 2006 M, Aunola S, Eriksson JG, Hemio K, Ha- ENDpoints and Outcomes. Clinical 14. Lebovitz HE: : a com- malainen H, Harkonen P, Keinanen-Kiu- Trials.gov Identifier: NCT00263042 mon link between type 2 diabetes and car- kaanniemi S, Laakso M, Louheranata A, (sanofi-aventis). http://clinicaltrials.gov/ diovascular disease. Diabetes Obes Metab Mannelin M, Paturi M, Sundvall J, Valle ct/show/NCT00263042 8:237–249, 2006 TT, Uusitupa M, Tuomilehto J, Finnish 35. Narayan KM, Hoskin M, Kozak D, Kriska 15. Bjo¨rntorp P: Body fat distribution, insulin Diabetes Prevention Study Group: Sus- AM, Hanson RL, Pettitt DJ, Nagi DK, Ben- resistance, and metabolic . Nutri- tained reduction in the incidence of type 2 nett PH, Knowler WC: Randomized clin- tion 13:795–803, 1997 diabetes by lifestyle intervention: fol- ical trial of lifestyle interventions in Pima 16. Bjo¨rntorp P: Regional fat distribution: im- low-up of the Finnish Diabetes Preven- Indians: a pilot study. Diabet Med 15:66– plications for type II diabetes. Int J Obes tion Study. Lancet 368:1673–1679, 2006 72, 1998 Relat Metab Disord 16 (Suppl. 4):S19– 26. Holt R: Orlistat reduces features of the 36. Mayer-Davis EJ, D’Antonio AM, Smith S27, 1992 metabolic syndrome: the XENDOS study. SM, Kirkner G, Levin Martin S, Parra-Me- 17. UK Prospective Diabetes Study (UKPDS) Diabetes Obes Metab 5:356, 2003 dina D, Schultz R: Pounds Off With Em- Group: Intensive blood-glucose control 27. Torgerson JS, Hauptman J, Boldrin MN, powerment (POWER): A clinical trial of with sulphonylureas or insulin compared Sjo¨stro¨m L: XENical in the prevention of strategies for black with conventional treatment and risk of diabetes in obese subjects (XENDOS) and white adults with diabetes who live in complications in patients with type 2 di- study: a randomized study of orlistat as an medically underserved rural communi- abetes (UKPDS 33). Lancet 352:837–853, adjunct to lifestyle changes for the pre- ties. Am J Public Health 94:1736–1742, 1998 vention of type 2 diabetes in obese pa- 2004 18. Yki-Ja¨rvinen H: Thiazolidinediones. tients. Diabetes Care 27:155–161, 2004 37. Ryan DH, Espeland MA, Foster GD, N Engl J Med 351:1106–1118, 2004 28. Chiasson JL, Josse RG, Gomis R, Hanefeld Haffner SM, Hubbard VS, Johnson KC, 19. Hedblad B, Nilsson PM, Zambanini A, M, Karasik A, Laakso M: STOP-NIDDM Kahn SE, Knowler WC, Yanovski SZ: Berglund G: Rosiglitazone reduces carotid Trail Research Group: Acarbose for pre- Look AHEAD Research Group: Look intima-media thickness progression over vention of type 2 diabetes mellitus: the AHEAD (Action for Health in Diabetes): 1 year in patients with type 2 diabetes but STOP-NIDDM randomised trial. Lancet design and methods for a clinical trial of not in non-diabetic people who have in- 359:2072–2077, 2002 weight loss for the prevention of cardio- sulin resistance alone. J Intern Med 261: 29. DREAM (Diabetes REduction Assessment vascular disease in type 2 diabetes. Con- 293–305, 2007 with ramipril and rosiglitazone Medica- trol Clin Trials 24:610–628, 2003

S282 DIABETES CARE, VOLUME 31, SUPPLEMENT 2, FEBRUARY 2008 Nilsson

38. Sjo¨stro¨m CD, Peltonen M, Wedel H, Sjo¨s- (Suppl. 1):S2–S4, 2001 Swedish Obese Subjects study: effects of tro¨m L: Differentiated long-term effects of 40. Sjo¨stro¨m L, Narbro K, Sjo¨stro¨m CD, Kara- bariatric surgery on mortality in Swedish intentional weight loss on diabetes and son K, Larsson B, Wedel H, Lystig T, Sul- obese subjects. N Engl J Med 357:741– hypertension. Hypertension 36:20–25, livan M, Bouchard C, Carlsson B, 752, 2007 2003 Bengtsson C, Dahlgren S, Gummesson A, 41. Fonarow GC: Beta-blockers for the post- 39. Torgerson JS, Sjo¨stro¨m L: The Swedish Jacobson P, Karlsson J, Lindroos AK, Lo¨n- myocardial infarction patient: current Obese Subjects (SOS) study: rationale and roth H, Na¨slund I, Olbers T, Stenlo¨f K, clinical evidence and practical consider- results. Int J Obes Relat Metab Disord 25 Torgerson J, Agren G, Carrlsson LM: ations. Rev Cardiovasc Med 7:1–9, 2006

DIABETES CARE, VOLUME 31, SUPPLEMENT 2, FEBRUARY 2008 S283