Paul Ernsberger, PhD, and Richard J. Koletsky, MD

To establish if obesity shortens life expectancy, it must first be shown that obese people consistently die sooner than appropriate controls. Evidence for a causal relationship must then be obtained. Alternatives to the hypothesis of simple causality must be refuted. Furthermore, reversing the condition should extend life.

The relationship between initial body weight, as measured by body mass index (BMI) and subsequent mortality, is complex and varies markedly between different prospective epidemiologic studies (Table 1) (1–20). Most studies have shown a U-shaped relationship between BMI and mortality, with both low and high body weights associated with increased risk of death (21). Others show no relationship (22, 23). A few studies show a steady decline in mortality with increasing BMI, and a few show a steady increase in mortality with increasing weight up to a relative risk (RR) between 1.5 and 2 for the extremely obese. The latter studies, listed in the top few rows of Table 1, have been emphasized in the setting of public health policy concerning body weight.

A meta-analysis of 23 major studies showed a U- shaped curve in the combined data for men and women, with increased mortality when BMI was under 23 or above 28 (24). To place this in perspective, a BMI of 21 corresponds to the old “ideal weight” set by Metropolitan Life Insurance. Thus, being less than 10 percent “overweight” or being underweight by any amount raises the risk of death. At the other end of the spectrum, being more than 35 percent “overweight” also constitutes a hazard.

Which is more important in the overall population, the hazards of underweight or the hazards of overweight? In a study of Finnish women aged 25 to 64, the leanest one-fifth and the fattest one-fifth of the population both showed a RR of about 1.5 (.7). For women over 65, only underweight was a hazard. Therefore, the impact of underweight and overweight are very nearly the same. Nearly equivalent results have been obtained in many populations whenever “U-shaped” distribution of mortality versus BMI has been obtained. Based on his study of 1.8 million Norwegians, Waaler estimated that 4.3 percent of deaths in persons aged 30 to 79 could be attributed to obesity (25) Especially if the elderly are included, the number of excess deaths attributable to underweight may be higher. A rational approach to public health would dictate devoting equal time to treating underweight as overweight..

Smoking and Wasting Disease

Some authors acknowledge that moderately fat people have increased life expectancy, but nonetheless argue that obesity is dangerous (26). It has been claimed that because cigarette smokers weigh an average of 2 to 3 kg less than nonsmokers, the excess mortality among thin people can be explained by increased smoking.  However, in numerous studies, mild obesity was found to be protective in nonsmokers and smokers alike. Adjusting the mortality data for smoking has no effect on its relationship to body weight. Importantly, only light-to-moderate cigarette smokers tend to be lean. In many studies, persons who consume at least two packs a day actually weigh more than nonsmokers. Thus, the smokers at the highest risk for disease are not unusually lean.

Another explanation offered for the increased mortality among underweight persons is that their thinness is caused by a hidden, undetectable disease that causes them to lose weight and then, after many years of thinness, causes their death. However, there is no evidence that “occult wasting disease” even exists. Patients with unexplained weight loss tend to either deteriorate or die within a single year or to regain their lost weight along with their health (27). Significant weight loss is a late symptom in cancer, appearing only after the tumor is large enough to have metabolic effects or cause pain. Because weight loss occurs only in the advanced stages of serious disease, nearly all persons who have become lean due to fatal wasting disease would be screened out of prospective studies during the initial physical examination. Any such persons accidentally included should die within the first 2 years of follow-up. However, in most studies, excluding deaths occurring in the first 5 to 10 years has no effect on the outcome.

Risk Changes with Age

Age has a powerful impact on the relationship between body weight and mortality. Studies of people between the ages of 25 and 45, such as first-time life insurance buyers and young nurses, tend to show the greatest hazard of obesity (see Table 1, top two lines). In contrast, among the elderly low body weight is a strong risk for subsequent death, but there is little hazard of obesity until extreme levels are reached (BMI > 35). Numerous studies have demonstrated that low body weight is a reliable harbinger of decline and death in persons over 60 years of age.

The impact of age on the relative hazard of obesity is illustrated in a prospective study of British civil servants (adjusted for smoking) (28). In the youngest age group, the mortality rate doubled from the lightest to the heaviest subgroup. This doubling of risk agrees with a report from the Nurses’ Health Study, which found an identical increase in women of about the same age (1). The Nurses’ Health Study reported its findings only as relative risk, thereby emphasizing the two-fold elevation in hazard. What was left unstated is the very low risk of dying at these young ages. At age 44 you have a 98 percent chance of living to age 54 if you are lean and 96 percent chance if you are severely obese.
What is considerably more important from a public health standpoint is the effect of weight on mortality at ages 60 to 64, where mortality is highest for those who meet life insurance criteria. If you are obese, you face a 2 percent greater chance of dying between 44 and 54 but a 6 percent lesser chance of dying between 64 and 74. As a result, the net adverse impact of obesity on median life expectancy is minimal to nonexistent.

Table 1. “Population characteristics and outcome of epidemiologic studies of obesity and total mortality.”

Study Population - Relative Hazard of Obesity - Likely Prevalence of Weight Loss Practices - Reference
Young nurses - Exceptionally high - Very high - (Manson et al., 1995)
Holders of individual life insurance policies - Very high - Very high - (Lew et al., 1979)
Harvard alumni - Very high - High - (Lee et al., 1993)
Residents of affluent Boston suburb (Framingham) - High - High - (Garrison et al., 1983)
Neighbors and relatives of American Cancer Society volunteers - High - High - (Lew et al., 1979)
Residents of Finland - Moderate - Moderate - (Rissanen et al., 1989; 1991)
White women in Charleston, SC - Moderate - Moderate - (Stevens et al., 1992)
Black women in Charleston, SC - None - Low - (Stevens et al., 1992)
Civil servants of rural eastern Finland - None - (women); low (men) - Low - (Tuomilehto et al., 1987)
German construction workers - Inverse - Low - (Brenner et al., 1997)
Dutch civil servants - None (women); low (men) - Low - (Tuomilehto et al., 1987)
Black Kaiser Permanente subscribers - None (women); U-shaped* (men) - Low - (Wienpahl et al., 1990)
Residents of villages in rural Italy - None - Very low - (Anonymous, 1982)
Residents of villages in rural Scotland - Inverse relation - Very low - (Garn et al., 1983)
San Francisco longshoremen - Inverse - Low - (Borhani et al., 1963)
Elderly populations - Inverse relation - Very low - See text discussion
Residents of American Samoa - None - Very low - (Crews, 1989)
Residents of Micronesia - Inverse relation - Very low - (Vandenbroucke et al., 1984)
Residents of Fiji - Inverse relation - Very low - (Hodge et al., 1996; Collins et al., 1998)
New Zealand Maori - None - Very low - (Salmond et al., 1985)
Native Americans of the Pima tribe - Inverse relation (women);inverse up to BMI of 40 (men) - Very low - (Hanson et al., 1995)

Studies are ranked according to the degree of relative risk associated with obesity. The incidence of weight loss behavior is based either on direct report or on the characteristics of populations known to be associated with a high prevalence of dieting practices.
*Highest mortality at body mass index (BMI less than 20); lowest mortality at BMI of 28 (moderately overweight). Risk increase only for BMI greater than 40. Reprinted with permission from Journal of Social Issues 1999; 55: 224–225.

Health Benefits of Obesity

The fact that obese persons have a normal life expectancy presents a paradox, since the incidence of a number of serious risk factors is increased in obesity. The solution to this puzzle is that there are advantages as well as disadvantages to being heavy (29). Obese persons are less likely to later develop cancer, as shown in studies in France, Hawaii, Iowa, Massachusetts, Norway, Puerto Rico, and Scotland, as well as the Hypertension Detection and Follow-up Program and Seven Countries studies. The obese are also protected against infectious diseases, chronic obstructive pulmonary disease, osteoporosis, mitral valve prolapse, intermittent claudication, renovascular hypertension, eclampsia, premature birth, anemia, diabetes type I, peptic ulcer, scoliosis, and suicide (29). These health benefits of obesity might potentially offset its hazards.

Obesity is also associated with improved survival in several diseases. Heavy persons with hypertension, diabetes type II, and high cholesterol have a more favorable prognosis than thin people with these same ailments (29). Obese hypertensives have been shown to outlive lean hypertensives in more than 15 separate controlled prospective studies (30). In one study, 43 percent of nonobese hypertensives died, whereas only 26 percent of obese hypertensives died. In the Systolic Hypertension in the Elderly Program, mortality fell by 35 percent for each 5 increments in BMI. Lean Type II diabetics tend to have a more severe form of the disease than obese diabetics (31).

Obesity may cushion the risk of high cholesterol. Heavy persons exhibit elevations in literally every proposed risk factor for atherosclerosis: total cholesterol, LDL/HDL ratio, triglycerides, uric acid, glucose, and insulin. Despite this, a comprehensive review of the medical literature came to the conclusion that obesity is not associated with plaque formation in blood vessels (32). Even among persons weighing over 300 pounds, no increase in coronary plaque has been found at autopsy. Although heavy people have elevated risk factors, they fail to translate into a higher incidence of disease.

Although hypertension, diabetes, and hyperlipidemia have reduced complications and mortality in heavy persons, this does not mean these conditions are benign in obesity, nor does it mean that people with diabetes and hypertension should be encouraged to gain weight, since this may worsen their condition. However, the threat to the health and longevity of fat people posed by diabetes and hypertension is overestimated, due to the failure to take into account the ameliorating influence of obesity on these conditions.

Increasing Weight, Declining Mortality

Epidemic increases in the incidence of obesity in the U.S. and around the world have been documented in numerous independent reports since the 1980s. Mortality rates have fallen simultaneously. The causes of death contributing to the declines in mortality are those usually linked to obesity, especially heart attack and stroke. If obesity were truly a major contributor to premature death, as some have claimed (1), then we would expect a rising death toll, especially with the sharp rise in obesity since 1977.

An example of short-term time trends in a specific population comes from Minneapolis-St. Paul. Between 1981 and 1986, the average BMI rose 1.2 in women and 0.6 in men (33). Yet, during the same time interval and in the same population, the incidence of coronary heart disease fell 13 percent in women and 20 percent in men. Paralleling the decrease in coronary disease were falls in risk factors: diastolic blood pressure (down 0.9 and 1.1 mm Hg in women and men, respectively) and cholesterol (down 5.8 and 5.2 mg/dl).

Even as human populations fatten, risk factors and diseases usually linked to obesity continue to decline. The most likely explanation is that improvements in lifestyle, such as reduced intakes of cholesterol, saturated fat, and sodium and increased exercise, have improved health but have not led to widespread loss of weight. Also, much of the decline in mortality rates in western cultures has been credited to improved detection and treatment of hypertension with improved medications. Hypertension is the primary health risk faced by the obese, and if adjustment is made for the higher blood pressures in obese persons, their elevated rates of heart attack and stroke are eliminated (22). Thus, improved treatment of hypertension in recent years may have eliminated much of the cardiovascular risk associated with obesity.

Successful Weight Programs

The only way to establish if obesity is a cause of premature mortality is to conduct a randomized trial in which one group of obese persons would undergo weight reduction while a control group received comparable counseling on nutrition and exercise without specific advice to lose weight. Any differences in disease incidence or mortality between the groups could then be attributed to the effect of weight loss. Unfortunately, such a trial is impossible because no safe and effective treatment for obesity exists.

On the other hand, successful weight gain programs have been shown to lower the risk of death in excessively lean persons. Weight gain is the preferred course of treatment and is life saving for both anorexia nervosa and protein-calorie malnutrition. Elderly patients enrolled in a weight gain program who increased their weight by at least 5 percent showed reduced mortality and extended longevity (34). In patients with lung disease having a BMI of 25 or less, a gain of 2 kg or more reduced mortality (35). Thus, controlled clinical trials of weight gain in underweight persons show decreased mortality, but no equivalent trial exists to show benefits of weight loss.

Weight and Mortality in Patients

It is people who are ill who are most likely to be counseled to lose weight. We should therefore explore the relationship between body weight and mortality in the unwell.

In hospitalized patients, BMI upon admission was compared to the odds of dying in the hospital after adjustment for disease severity (36). The highest risk of death was in people who were underweight (BMI less than 21). At all ages, the risk of death declined with increasing overweight. For patients aged 50 to 79, the best chance of survival was at a BMI of 40, usually considered “morbidly obese.” For patients in their 80s, the optimum BMI is around 32, which is considered “severely obese.” The authors of this study considered whether the excess mortality in nonobese patients was caused by malnutrition due to chronic illness. To test this, they examined weight change between hospitalizations for those who were hospitalized more than once. There was no increase in risk of death for those who were losing weight prior to hospitalization, indicating that it is long-term, possibly lifelong, leanness that raises the risk of in-hospital death.

These findings have been confirmed in other studies. The Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) found a lack of overweight (BMI

< 25) to be an independent predictor of mortality in hospitalized patients after stratified adjustment for 15 different physiologic and demographic variables (37). Thus, even though the absence of obesity may be favorable for young and healthy individuals, these results are consistent with the notion that obesity facilitates survival when the individual is challenged by illness or aging.

Weight Cycling and Obesity-Associated Illness

A further difference between the populations at the top and the bottom of Table 1 is the likely prevalence of dieting and participation in weight loss programs. High socioeconomic status is associated with increased efforts to lose weight. The relative hazard of obesity is increased in younger age groups, which are also most likely to pursue weight loss. Weight loss methods may be hazardous in themselves. Nurses in particular are exposed to opportunities for pharmaceutical, surgical, and very low calorie diet interventions, which can be hazardous.

There is also strong evidence that the regain of weight that almost always follows weight loss can be harmful. Thus, the harmful effects of weight cycling may contribute to the high mortality risk of obesity in populations where weight loss is prevalent, in contrast to the low mortality risk in populations where obesity is the norm and likely to be accepted or tolerated.

Fatal consequences of refeeding after starvation are well known in famine victims and anorexics. Obese dieters may not be immune from the adverse effects of refeeding after significant weight loss. Obese humans typically show repeated loss and regain of large amounts of weight. Men with large fluctuations in weight between the ages of 20 and 40 exhibit increased systolic and diastolic blood pressure and cholesterol (38). These “yo-yo dieters” are twice as likely to die of coronary heart disease, even after adjustment for known risk factors, than are men with stable or steadily increasing weight (38, 39). Fluctuations of body weight have been shown in many other major epidemiologic studies to have deleterious cardiovascular effects resulting in increased mortality.

A few studies have not found that weight cycling increases the risk of mortality. However, the populations lacking a risk associated with weight cycling tend to be those with low rates of dieting for weight loss, such as a study of elderly men in Baltimore (40). The lack of an effect of weight cycling in a population with a low level of intentional weight loss implies that minor random fluctuations in weight do not pose a health hazard.

In the Framingham study, the increase in cardiovascular disease in obese patients could be entirely explained by taking two facts into account (39). First, obese people were more likely to go up and down in weight than thin people. Second, weight cycling was associated with increased rates of death from cardiovascular disease. Obese people who maintained a high but steady weight had only an average risk of death or cardiovascular disease. This finding raises the possibility that much of the increased risk of disease and death in obese people is the result of repeated cycles of weight loss and regain.

Promote Healthy Lifestyles

This review of diabetes, hypertension, and hypercholesterolemia (Part 1) and mortality (Part 2) leads us to conclude that the money and effort spent on the generally unsuccessful pursuit of thinness would be better used directly promoting healthier lifestyles.



Paul Ernsberger, PhD, is an associate professor in nutrition, medicine (hypertension), pharmacology, and neuroscience at Case Western Reserve University School of Medicine in Cleveland, Ohio. With Jeanine Cogan, he coedited “Dying to be thin in the name of health: shifting the paradigm,” a special issue of the Journal of Social Issues, Volume 55, 1999. Richard J. Koletsky, MD, is an assistant clinical professor of medicine, the director of the Cleveland Weight Wellness Center, and does research at Case Western Reserve University.



References
1. Manson JE, Willett WC, Stampfer MJ, et al. Body weight and mortality among women. N Engl J Med 1995; 333: 677–685.

2. Lew EA, End JA, Wilber JA. The new build and blood pressure study. Trans Assoc Life Insur Med Dir Am 1979; 62:154–174.

3. Lee IM, Manson JE, Hennekens CH, et al. Body weight and mortality. A 27-year follow-up of middle-aged men. JAMA 1993; 270:2823–2828.

4. Garrison RJ, Feinleib M, Castelli WP, et al. Cigarette smoking as a confounder of the relationship between relative weight and long-term mortality. The Framingham Heart Study. JAMA 1983; 249:2199–2203.

5. Lew EA, Garfinkel L. Variations in mortality by weight among 750,000 men and women. J Chronic Dis 1979; 32:563–576.

6. Rissanen A, Heliovaara M, Knekt P, et al. Weight and mor¬tality in Finnish men. J Clin Epidemiol 1989; 42:781–789.

7. Rissanen A, Knekt P, Heliovaara M, et al. Weight and mortality in Finnish women. J Clin Epidemiol 1991; 44: 787–795.

8. Stevens J, Keil JE, Rust PF, et al. Body mass index and body girths as predictors of mortality in black and white women. Arch Intern Med 1992; 152:1257–1262.

9. Tuomilehto J, Salonen JT, Marti B, et al. Body weight and risk of myocardial infraction and death in the adult population of eastern Finland. BMJ 1987; 295(6599):623–627.

10. Brenner H, Arndt V, Rothenbacher D, et al. Body weight, pre-existing disease, and all-cause mortality in a cohort of male employees in the German construction industry. J Clin Epidemiol 1997; 50:1099–1106.

11. Wienpahl J, Ragland DR, Sidney S. Body mass index and 15-year mortality in a cohort of black men and women. J Clin Epidemiol 1990; 43:949–960.

12. Borhani NO, Hechter HH, Breslow L. Report of a 10- year follow-up study of the San Francisco longshoremen. J Chronic Dis 1963; 16:1251–1266.

13. Incidence and prediction of coronary heart disease in two Italian rural population samples followed-up for 20 years. Acta Cardiol 1982; 37:129–145.

14. Garn SM, Hawthorne VM, Pilkington JJ, et al. Fatness and mortality in the West of Scotland. Am J Clin Nutr 1983; 38:313–319.

15. Crews DE. Multivariate prediction of total and cardiovascular mortality in an obese Polynesian population. Am J Public Health 1989; 79:982–986.

16. Vandenbroucke JP, Mauritz BJ, de Bruin A, et al. Weight, smoking, and mortality. JAMA 1984; 252:2859–2860.

17. Hodge AM, Dowse GK, Collins VR, et al. Mortality in Micronesian Nauruans and Melanesian and Indian Fijians is not associated with obesity. Am J Epidemiol 1996; 143:442–455.

18. Collins VR, Dowse GK, Cabealawa S, et al. High mortality from cardiovascular disease and analysis of risk factors in Indian and Melanisian Fijians. Int J Epidemiol 1998; 25:59–69.

19. Salmond CE, Beaglehole R, Prior IA. Are low cholesterol values associated with excess mortality? BMJ 1985; 290 ( 6466):422–424.

20. Hanson RL, McCance DR, Jacobsson LTH, et al. The U- shaped association between body mass index and mortality: relationship with weight gain in a native American population. J Clin Epidemiol 1995; 48:903–916.

21. Andres R. Effect of obesity on total mortality. Int J Obes 1980; 4:381–386.

22. Keys A. Overweight, obesity, coronary heart disease, and mortality: the W.O. Atwater Memorial Lecture, 1980. Prog Clin Biol Res 1981; 67:31–46.

23. Keys A. Longevity of man: relative weight and fatness in middle age. Ann Med 1989; 21:163–168.

24. Troiano RP, Frongillo EA Jr, Sobal J, et al. The relationship between body weight and mortality: a quantitative analysis of combined information from existing studies. Int J Obes 1996; 20:63–75.

25. Waaler HT. Height, weight, and mortality. The Norwegian experience. Acta Med Scand Suppl 1984; 679:1–56.

26. Lee IM, Manson JE. Body weight and mortality: what is the shape of the curve? Epidemiology 1998; 9:227–228.

27. Wallace JI, Schwartz RS, LaCroix AZ, et al. Involuntary weight loss in older outpatients: incidence and clinical significance. J Am Geriatr Soc 1995; 43:329–337.

28. Jarrett RJ. Is there an ideal body weight? BMJ 1986; 293(6545):493–495.

29. Ernsberger P, Haskew P. Health implications of obesity: an alternative view. Journal of Obesity and Weight Regu¬lation 1987; 6:55–137.

30. Barrett-Connor E, Khaw KT. Is hypertension more benign when associated with obesity? Circulation 1985; 72:53–60.

31. Ross C, Langer RD, Barrett-Connor E. Given diabetes, is fat better than thin? Diabetes Care 1997; 20:650–652.

32. Barrett-Connor EL. Obesity, atherosclerosis, and coronary artery disease. Ann Intern Med 1985; 103(6 [Pt 2]): 1010–1019.

33. Sprafka JM, Burke GL, Folsom AR, et al. Continued decline in cardiovascular disease risk factors: results of the Minnesota Heart Survey, 1980–1982 and 1985–1987. Am J Epidemiol 1990; 132:489–500.

34. Keller HH. Weight gain impacts morbidity and mortality in institutionalized older persons. J Am Geriatr Soc 1995; 43:165–169.

35. Schols AM, Slangen J, Volovics L, et al. Weight loss is a reversible factor in the prognosis of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998; 157(6 [Pt 1]):1791–1797.

36. Potter JF, Schafer DF, Bohi RL. In-hospital mortality as a function of body mass index: an age-dependent variable. J Gerontol 1988; 43(3):59–63.

37. Galanos AN, Pieper CF, Kussin PS, et al. Relationship of body mass index to subsequent mortality among seriously ill hospitalized patients. Crit Care Med 1997; 25: 1962–1968.

38. Hamm P, Shekelle RB, Stamler J. Large fluctuations in body weight during young adulthood and twenty-five- year risk of coronary death in men. Am J Epidemiol 1989; 129:312–318.

39. Lissner L, Odell PM, D’Agostino RB, et al. Variability of body weight and health outcomes in the Framingham population. N Engl J Med 1991; 324:1839–1844.

40. Lissner L, Andres R, Muller DC, et al. Body weight variability in men: metabolic rate, health and longevity. Int J Obes 1990; 14:373–383.

This article is from the Healthy Weight Journal (Health at Every Size Journal) and can be cited as Paul Ernsberger, PhD, and Richard J. Koletsky, MD, “Part 2: Rationale for a Wellness Approach to Obesity” from Health At Every Size 14:2 (March/April 2000).