The Obesity Paradox and the Flegal JAMA Study - Fact or Fiction?
Critical Appraisal of the Obesity Paradox and the Flegal Study (JAMA, Jan 2013)
Director of Medical Research, No Surrender Breast Cancer Foundation (NSBCF)
European Association for Cancer Research (EACR)
Where Is Sanity?
Recently we have been treated yet again to the usual cacophony of poorly informed and misleading headlines, this time concerning the "obesity paradox" - that overweight and low-grade (grade 1) obesity are not associated with compromised mortality and that being overweight in fact is associated with lower mortality compared to normal weight people. So from the news media we have, with the NY Times Health section leading the pack in unsurprisingly hyperbolic fashion: "Study Suggests Lower Mortality Risk for People Deemed to Be Overweight" (NY Times, 1 Jan 2013), and even more recklessly from the same source "Our Absurd Fear of Fat" (NY Times, 2 Jan 2013), joining these others: "Big deal: You can be fat and fit" (CNN, 3 Jan 2013), "Being overweight linked to lower risk of mortality" (CNN International, 2 Jan 2013), "Few Extra Pounds Won't Kill You—Really" (WSJ, 1 Jan 2013), "Being moderately overweight might not pose health risk" (LA Times, 1 Jan 2013), "Being overweight may increase odds of living longer" (Fox News, 2 Jan 2013), and from so-called "health" e-zines we have "Why Do Fat Guys Live Longer?" (Men's Health, 2 Jan 2013), and from even medical news sources we have "A Few Extra Pounds Linked to a Longer Life" (WebMD, 1 Jan 2013), "A Bit Of Extra Weight Helps You Live Longer" (Medical News Today, 2 Jan 2013), and with even the respected NIH's MedlinePlus weighing in with "More evidence for "obesity paradox"" (1 Jan 2013). Listening to these misinformed voices, as a medical researcher I bemoan the lost art of critical appraisal and the negative effects of the false assurance these headlines, and the study underlying them, will inevitably have. Let's critically review the claims and see what's really going on.
Flaws and Limitations, and Arguing Against the Facts
There are several significant methodological issues with the widely cited and - we will show widely misinterpreted - systematic review and meta-analysis (SR/MA) conducted recently by Katherine Flegal and colleagues  at the Centers for Disease Control and Prevention on the association between overweight and obesity, and all-cause mortality using standard categories of BMI (Body Mass Index). First and foremost is the choice of BMI as weight metric. In the Iowa Women's Health Study , compared to BMI waist circumference (WC) was demonstrated to be superior as a risk indicator for all-cause mortality. Whereas BMI was associated with mortality in a J-shaped fashion, with mortality rates being elevated in the leanest as well as in the most obese women, the waist/hip circumference ratio was strongly and positively associated with mortality in a strictly dose-response manner. Indeed, clinical guidelines from the Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults  have recommended that when assessing risk of adiposity-related disease, both BMI and waist circumference should be considered, and the unabated increases in waist circumference (WC) and its the superior association with obesity are now extensively evidenced (see  and references therein).
In this same connection, the Danish Cancer Society (DCS) study from Jane Bigaard and colleagues  found that a 10% larger waist circumference corresponded to a 1.48 times higher mortality over the entire range of waist circumference, with a dose-response relationship between waist circumference and mortality for individuals of the same BMI throughout the spectrum (even for those in the normal-weight range as judged by BMI). Similarly, a meta-analysis  of 29 elderly (65 - 74 yrs) cohorts found that a large WC, defined as ≥102 cm (40.2 inches) in men and ≥88 cm (34.6") in women, was associated with increased all-cause mortality relative risks for those the overweight and obese BMI categories compared with the ‘healthy’ weight and a small WC (<94 cm (37") in men; <80 cm (31.5") in women) category, showing that in and so in elderly people with an increased WC there were increased mortality risks - even across BMI categories, and for those who were classified as ‘underweight’ using BMI (see also the REGARDS Study ).
These studies collectively contradict the findings of the Flegal SR/MA  as to significantly higher all-cause mortality RR in overweight and obese (grades 2 and 3) BMI categories relative to normal weight, and further demonstrate the inadequacy of BMI alone as a measure of adiposity.
In addition, in the general population abdominal obesity (waist circumference >88 cm in women and >102 cm in men) in adults with BMI <35 kg/m2 was associated strongly with multiple cardiovascular risk factors (CRF), these including hypertension, diabetes, and hyperlipidemia [7, 8], suggesting the limitation of Flegal study in its restriction solely to all-cause rather than cause-specific mortality. This is further suggested in the study from Stephen Farrell and colleagues  who demonstrated that low cardiorespiratory risk fitness (CRF) is a stronger predictor than BMI of all-cause mortality in women, with low CRF in women being an important predictor of all-cause mortality, again suggesting that BMI as a predictor of all-cause mortality risk in women can be misleading unless CRF is also evaluated and weighted in.
An additional methodological limitation is failure to control for the aggregate effects of smoking and reverse causality, that latter being the state in which diseases lead to both weight loss and higher mortality. The Scottish study from Debbie Lawlor and colleagues  that examined the unbiased association of directly measured overweight and obesity (based on direct measurement of weight and height, not BMI) with all-cause and cause-specific mortality in two large prospective cohort studies. With the first 5 years of deaths removed as control, overweight was associated with an increase in all-cause mortality among never-smokers (relative risk range = 1.12 to 1.38), while obesity was associated with a doubling of risk in men in both cohorts and a 60% increase in women. Furthermore, in both never-smokers and current smokers, being overweight or obese was associated with significant increase in risk of cardiovascular disease.
These findings demonstrate that with appropriate control for smoking and reverse causality, both overweight and obesity are associated with important increases in all-cause and cause-specific mortality, and in particular with cardiovascular disease mortality. What the Scottish study reveals uniquely is that smoking is in fact itself associated with lower BMI and is moreover strongly associated with many of the same adverse health outcomes such as diabetes, cardiovascular disease, and respiratory disease that are likely to show increased incidence in the overweight and obese, and finally, given the strong association of smoking with reduced BMI, it may be insufficient of wholly control its effects through the typical simple adjustment for smoking in multivariable models, since for instance measurement error is effectively inevitable (current smokers may for example identify themselves as past smokers). In addition, this Scottish study shows that reverse causality - in which diseases lead to both weight loss and higher mortality - could attenuate the apparent relationship of obesity to mortality, especially but not exclusively in studies with shorter follow-up periods and/or inadequate control for critical confounders such as smoking.
The Scottish study  therefore demonstrates that reverse causality may induce an appreciable underestimation of the effect of obesity on all-cause mortality , while smoking strongly works to effectively mask the effects of both overweight and obesity on all-cause and cause-specific mortality, consequent to both the strong association between smoking and lower BMI, and to the strong effect of smoking on all-cause, and cause-specific cardiovascular, and cancer mortality, since mortality in those with lower BMI is increased because of their greater likelihood of being smokers, not because lower BMI is inconsequential to health. And being either overweight or obese was associated with significant increases in cardiovascular (particularly from coronary heart disease) and in oncological mortality among never-smokers, and even in current smokers, being overweight or obese was still associated with insignificant increase in risk of cardiovascular disease. It is simply that smoking and reverse causality alone and more powerfully, jointly, mask the true effects of overweight and obesity on all-cause and cause-specific mortality.
This also underlines the highly questionable constraint to all-cause mortality adopted by the Flegal study , rather than assessment of cause-specific mortality from cancer, heart disease or diabetes, given that any association between weight and mortality for different disease categories may vary with the specific disease, and moreover may show stronger linkage with weight at lower thresholds of BMI than does all-cause mortality, not of course neglecting the clear importance of morbidity and disability and compromised QoL associated with long-term diseases.
It also neglects to caution the reader that BMI is not itself decisively established as a reliable measure of unhealthy adiposity (‘fatness’) since it is only metric of height and hence fails to account for other known disease and mortality factors such as (1) differing fat levels, or (2) differing fat distribution such as the highly unfavorable "sarcopenic obesity" in which we have elevated fat mass concurrently with lowered lean / muscle mass), or (3) muscularity, or (4) nutritional balance, among others (remember highly muscular individual for example can have a high BMI and therefore be categorized as overweight, while not necessarily carrying significant excess fat). In addition, while BMI reflects the influence of body height over body weight, it does not reveal body fat percentage (BFP), and it has been established that body fat percentage (BFP) correlates with risk factors for cardiovascular disease and metabolic syndrome and hence may be a useful predictor of risk, particularly in metabolically obese, normal weight individuals . It is clear that focusing of the narrow associative "play" between BMI and mortality (and more artificially still, only all-cause mortality) is unrepresentative of real world practice where a wide spectrum of risk factors would assuredly be weighed to assess mortality (and morbidity) risk, such as hypertension, dyslipidemia, and glucose and insulin dysfunction among many others, with BMI playing a highly limited role in whole- patient risk evaluations. And it is now clear that abdominal obesity is more harmful than general obesity: so, for example the risk of diabetes increases with increases in abdominal fat mass, waist circumference, or waist-to-hip circumference ratio, and this is in fact independent of BMI value [17. 18], while viscerally deposit fat as opposed to fat elsewhere in the body is associated with higher risk for hypertension , and abdominal fat mass is a strong risk factor for stroke independent, gain independently of BMI .
But still further and critical erosion of the Flegel study's conclusions come from the recent NCI prospective cohort study from Yikyung Park and NCI co-researchers  who performed a prospective analysis using data from 183,211 adults aged 45–75 who enrolled in the population-based Multiethnic Cohort Study. This NCI review exercised uncommonly effective control for confounding from conditions that lead to weight loss and mortality by excluding participants (1) with a history of cancer or heart disease, (2) who ever smoked, and (3) who died within the first 3 years of follow-up. Under those controls, an increased risk of mortality was observed in participants with a BMI ≥ 27.5 in both men and women compared with the reference category of BMI 23.0–24.9 (with a BMI ≥ 35.0 carrying a greater risk of mortality in men than in women), so that among healthy never smokers, adult overweight and obesity were both associated with increased risk of mortality in both genders, confirming the findings of the other tightly controlled study just discussed, the Scottish study.
Therefore, the balance of critically reviewed and appraised evidence fails to support the contention of the Flegal study that overweight is associated with significantly lower all-cause mortality relative to normal weight and that obesity at certain lower levels (grade 1) is not associated with higher mortality, but rather the cumulative weight of the data to date continue to support the negative impact on health, mortality, and morbidity of being overweight or obese. And in this connection, it must be also remembered that intentional weight loss does in fact result in a decreased incidence of cancer, particularly female obesity-related cancers  and so the seductive but misleading findings of the Flegal study should not lull us into a false security and diminish the very real motivations for maintenance of health-favorable weight along with other dietary and lifestyle interventions.
The Obesity, Diabesity and Metabolic Syndrome Pandemics
With the prevalence of obesity being greater than 20% in many developed countries and increasing in developing countries, and with obesity being unambiguously associated with metabolic disorders, especially diabetes, cardiovascular diseases, pulmonary diseases, digestive diseases, and cancers, it is clear that weight loss must remain the central defense against the pandemic, associated with an increased risk of death, morbidity, and accelerated aging , of obesity , diabesity (the designation for the continuum of abnormal metabolic biologies from mild insulin resistance to established diabetes that includes any insulin dysfunctions secondary to obesity), and metabolic syndrome, and no one should, through the compromised methodologies and against-the-weight-of-the-evidence conclusions we have demonstrated within the Flegal study, sustain the illusion that overweight or obesity is other than detrimental and contributory to morbidity and mortality, both all-cause and cause-specific, within the context of these rising and unabated pandemics. We need such clear recognition to help avoid or mitigate what is expected to be the future obesity for adults in the United States, namely that by 2030 (as projected from the National Health and Nutrition Examination Study (NHANES)), 86.3% American adults will be overweight or obese, and 51.1% of them will be obese, with total health-care costs attributable to obesity/overweight doubling each decade to 2030 and accounting for 16–18% of total US health-care costs . And overweight and obesity could account for 14% of all deaths from cancer in men and 20% of deaths in women as shown a prospective study of 900,000 US adults . Therefore, as these considerations show, both overweight and obesity cause huge burdens for patients, family, and for society, locally and globally and cannot legitimately be perceived as in any way healthful. It will take a wide spectrum of coordinated health professionals, public education, sophisticated research into diagnostic and therapeutic tools and interventions, and proactive national policies to launch integrative and targeted initiatives to materially influence and slow these pandemics in obesity, diabesity and metabolic syndrome.
1. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA 2013 Jan 2; 309(1):71-82.
2. Folsom, A. R., Kaye, S. A., Sellers, T. A., et al (1993) Body fat distribution and 5-year risk of death in older women. JAMA. 269: 483–487.
3. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults—the evidence report. National Institutes of Health. ObesRes. 1998;6(suppl 2):51S-209S.
4. Bigaard J, Tjønneland A, Thomsen BL, et al. Waist circumference, BMI, smoking, and mortality in middle-aged men and women. Obes Res. 2003 Jul;11(7):895-903.
5. de Hollander EL, Bemelmans WJ, Boshuizen HC, et al. The association between waist circumference and risk of mortality considering body mass index in 65- to 74-year-olds: a meta-analysis of 29 cohorts involving more than 58 000 elderly persons. Int J Epidemiol 2012; 41(3):805-17.
6. Kramer H, Shoham D, McClure LA, et al. Association of waist circumference and body mass index with all-cause mortality in CKD: The REGARDS (Reasons for Geographic and Racial Differences in Stroke) Study. Am J Kidney Dis 2011; 58(2):177-85.
7. Janssen I, Katzmarzyk PT, Ross R. Body mass index, waist circumference, and health risk: evidence in support of current National Institutes of Health guidelines. Arch Intern Med. 2002;162(18):2074–2079. 8. Balkau B, Deanfield JE, Despres JP, et al. International Day for the Evaluation of Abdominal Obesity (IDEA): a study of waist circumference, cardiovascular disease, and diabetes mellitus in 168,000 primary care patients in 63 countries. Circulation. 2007;116(17):1942–1951.
9. Farrell SW, Braun L, Barlow CE, et al. The Relation of Body Mass Index, Cardiorespiratory Fitness, and All-Cause Mortality in Women. Obes Res 2002; 10(6):417-23.
10. Lawlor DA, Hart CL, Hole DJ, Davey Smith G. Reverse causality and confounding and the associations of overweight and obesity with mortality. Obesity (Silver Spring) 2006; 14(12):2294-304.
11. Park SY, Wilkens LR, Murphy SP, Monroe KR, Henderson BE, et al. (2012) Body mass index and mortality in an ethnically diverse population: the Multiethnic Cohort Study. Eur J Epidemiol 27: 489–497.
12. Birks S, Peeters A, Backholer K, et al. A systematic review of the impact of weight loss on cancer incidence and mortality. Obes Rev 2012; 13(10):868-91.
13. Lawlor (see below); Flanders WD, Augestad LB. Adjusting for reverse causality in the relationship between obesity and mortality. Int J Obes. 2008;32(Suppl 3):S42–6.
14. Chuang HH, Li WC, Sheu BF, et al. Correlation between body composition and risk factors for cardiovascular disease and metabolic syndrome. Biofactors 2012 Jul-Aug; 38(4):284-91.
15. Xiao J, Yang W. Weight loss is still an essential intervention in obesity and its complications: a review. J Obes 2012; 2012:369097.
16. Y. Wang, M. A. Beydoun, L. Liang, B. Caballero, and S. K. Kumanyika, “Will all Americans become overweight or obese? Estimating the progression and cost of the US obesity epidemic,” Obesity, vol. 16, no. 10, pp. 2323–2330, 2008.
17. H. Lundgren, C. Bengtsson, G. Blohme, L. Lapidus, and L. Sjostrom, “Adiposity and adipose tissue distribution in relation to incidence of diabetes in women: results from a prospective population study in Gothenburg, Sweden,” International Journal of Obesity, vol. 13, no. 4, pp. 413–423, 1989.
18. Q. Qiao and R. Nyamdorj, “Is the association of type II diabetes with waist circumference or waist-to-hip ratio stronger than that with body mass index,” European Journal of Clinical Nutrition, vol. 64, no. 1, pp. 30–34, 2010.
19. Z. Pausova, A. Mahboubi, M. Abrahamowicz, et al., “Sex differences in the contributions of visceral and total body fat to blood pressure in adolescence,” Hypertension, vol. 59, pp. 572–579, 2012.
20. F. Toss, P. Wiklund, P. W. Franks et al., “Abdominal and gynoid adiposity and the risk of stroke,” International Journal of Obesity, vol. 35, no. 11, pp. 1427–1432, 2011.
21. E. E. Calle, C. Rodriguez, K. Walker-Thurmond, and M. J. Thun, “Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. Adults,” The New England Journal of Medicine, vol. 348, no. 17, pp. 1625–1638, 2003.
22. Adams KF, Schatzkin A, Harris TB, et al. Overweight, obesity, and mortality in a large prospective cohort of persons 50 to 71 years old. N Engl J Med 2006;355:763-778.
23. Roth J, Qiang X, Marbán SL, et al. The obesity pandemic: where have we been and where are we going? Obes Res 2004; 12 Suppl 2:88S-101S.
Breast Cancer Watch