Advertisement

The Obesity Paradox and Weight Loss

      Abstract

      Background

      An “obesity paradox,” in which overweight and obese individuals with established cardiovascular disease have a better prognosis than normal weight subjects, has been reported in a number of clinical cohorts, but little is known about the effects of weight loss on the obesity paradox and its association with health outcomes.

      Methods

      Weight was determined in 3834 men at the time of a clinically referred exercise test and again during a clinical evaluation a mean of 7 years later. The associations among weight changes, baseline fitness, and other risk markers with cardiovascular and all-cause mortality were determined by Cox proportional hazards analysis.

      Results

      During the follow-up period, 314 subjects died (72 of cardiovascular causes). In a multivariate analysis (including baseline weight, weight change, exercise capacity, and cardiovascular disease), weight gain was associated with lower mortality and weight loss was associated with higher mortality (4% higher per pound lost per year, P<.001) compared with stable weight. For all-cause mortality, the relative risks for the no change, weight gain, and weight loss groups were 1.0 (referent), 0.64 (95% confidence interval, 0.50-0.83), and 1.49 (95% confidence interval, 1.17-1.89), respectively (P<.001). Those who died and exhibited weight loss had a significantly higher prevalence of deaths due to cancer and cardiovascular causes.

      Conclusion

      Weight loss was related to higher mortality and weight gain was related to lower mortality when compared with stable weight. The obesity paradox in our sample is explained in part by a combination of non-volitional weight loss related to occult disease and a protective effect of weight gain.

      Keywords

      SEE RELATED EDITORIAL p. 891
      The prevalence of obesity has increased dramatically over the last 2 decades and has reached global epidemic proportions in both adults and children in the United States.
      • Poirier P.
      • Giles T.D.
      • Bray G.A.
      • et al.
      Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      It is associated with numerous comorbidities, including hypertension, type 2 diabetes, dyslipidemia, and cardiovascular disease.
      • Poirier P.
      • Giles T.D.
      • Bray G.A.
      • et al.
      Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Klein S.
      • Burke L.E.
      • Bray G.A.
      • et al.
      American Heart Association Council on Nutrition, Physical Activity, and Metabolism Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism.
      Obesity is considered 1 of the 5 major risk factors for cardiovascular disease by the American Heart Association and has been shown to be associated with the development of heart failure, coronary heart disease, sudden cardiac death, atrial fibrillation, and reduced overall survival.
      • Poirier P.
      • Giles T.D.
      • Bray G.A.
      • et al.
      Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Klein S.
      • Burke L.E.
      • Bray G.A.
      • et al.
      American Heart Association Council on Nutrition, Physical Activity, and Metabolism Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism.
      Despite the associations between obesity and these health conditions, a number of recent studies have documented a puzzling phenomenon, termed the “obesity paradox,” in which overweight and obese individuals with established cardiovascular disease have a better prognosis compared with normal weight subjects.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Arena R.
      • Lavie C.J.
      The obesity paradox and outcome in heart failure: is excess bodyweight truly protective?.
      • Oreopoulos A.
      • Padwal R.
      • Kalantar-Zadeh K.
      • Fonarow G.C.
      • Norris C.M.
      • McAlister F.A.
      Body mass index and mortality in heart failure: a meta-analysis.
      Other recent studies have explored the role of fitness in the context of the obesity paradox and health outcomes. It has been demonstrated that irrespective of the presence of an obesity paradox or related risk markers, individuals who are relatively fit have significantly better survival and fewer cardiovascular events than unfit individuals.
      • McAuley P.
      • Myers J.
      • Abella J.
      • Froelicher V.
      Body mass, fitness and survival in veteran patients: another obesity paradox?.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      • McAuley P.
      • Pittsley J.
      • Myers J.
      • Abella J.
      • Froelicher V.F.
      Fitness and fatness as mortality predictors in healthy older men: the Veterans Exercise Testing Study.
      • Paradoxically, weight loss was related to higher mortality and weight gain was related to lower mortality when compared with stable weight over 7 years. Approximately 60% of the deaths in the weight loss group were attributable to conditions associated with muscle wasting.
      • The obesity paradox seems to be explained in part by a combination of non-volitional weight loss related to occult disease and a protective effect of weight gain.
      Although weight reduction is associated with a lessening of hypertension, diabetes, and other cardiometabolic risk markers,
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Klein S.
      • Allison D.B.
      • Heymsfield S.B.
      • et al.
      Waist Circumference and Cardiometabolic Risk: a Consensus Statement from Shaping America's Health: Association for Weight Management and Obesity Prevention; NAASO, the Obesity Society; the American Society for Nutrition; and the American Diabetes Association.
      there is controversy regarding whether weight loss decreases or increases mortality risk.
      • Lee I.M.
      • Paffenbarger Jr, R.S.
      Is weight loss hazardous?.
      • Sorensen T.I.A.
      Weight loss causes increased mortality: pros.
      • Yang D.
      • Fontaine K.R.
      • Wang C.
      • Allison D.B.
      Weight loss causes increased mortality: cons.
      Several recent observational studies have reported higher mortality risk with increased weight loss.
      • Myrskyla M.
      • Chang V.W.
      Weight change, initial BMI, and mortality among middle- and older-aged adults.
      • Østergaard J.N.
      • Grønbaek M.
      • Schnohr P.
      • Sørensen T.I.
      • Heitmann B.L.
      Combined effects of weight loss and physical activity on all-cause mortality of overweight men and women.
      • Ingram D.D.
      • Mussolino M.E.
      Weight loss from maximum body weight and mortality: the Third National Health and Nutrition Examination Survey Linked Mortality File.
      No doubt this issue is made more complex by the many comorbidities associated with obesity and non-volitional weight loss. Variations in preexisting conditions, subclinical disease, smoking, differences in the type of fat loss or fat distribution (eg, visceral fat vs lean mass), and the definition and measure used to classify obesity have been suggested to explain the inconsistent findings on weight loss and mortality.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Ross R.
      • Janiszewski P.M.
      Is weight loss the optimal target for obesity-related cardiovascular disease risk reduction?.
      We recently reported the existence of an obesity paradox in our population of veterans who were referred for exercise testing for clinical reasons.
      • McAuley P.
      • Myers J.
      • Abella J.
      • Froelicher V.
      Body mass, fitness and survival in veteran patients: another obesity paradox?.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      • Lissin L.W.
      • Gauri A.J.
      • Froelicher V.F.
      • Ghayoumi A.
      • Myers J.
      • Giacommini J.
      The prognostic value of body mass index and standard exercise testing in male veterans with congestive heart failure.
      We observed that exercise capacity strongly affects the obesity paradox; in fact, when stratified by fitness level, the paradoxic association between obesity and mortality was not found in the group registering high fitness.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      • McAuley P.
      • Pittsley J.
      • Myers J.
      • Abella J.
      • Froelicher V.F.
      Fitness and fatness as mortality predictors in healthy older men: the Veterans Exercise Testing Study.
      Stated differently, survival was influenced more by exercise capacity than body mass index. Given this finding and the controversial results of previous studies on the effects of weight loss on survival, it would be informative to document the effects of serial changes in both fitness and weight. In the current study, we assessed cardiovascular and all-cause mortality in a subgroup of patients in the Veterans Exercise Testing Study (VETS) who had weights recorded on 2 clinic visits before and after a treadmill test, a mean of 7 years apart. Our objectives were to determine the influence of fitness and changes in weight on cardiovascular and all-cause mortality and the extent to which changes in weight could explain the obesity paradox.

      Materials and Methods

      The VETS cohort is an ongoing, prospective evaluation of veteran subjects referred for exercise testing for clinical reasons, designed to address exercise test, clinical, and lifestyle factors and their association with health outcomes. From the VETS cohort, a list of approximately 10,000 male veterans who had undergone a maximal treadmill test for clinical reasons at the Veterans' Administration (VA) Palo Alto Health Care System between 1992 and 2008 for clinical indications was formed. This list was used to query the VA computerized database to identify patients with weights entered from clinic visits before and after an exercise test. The records were scanned to exclude patients with weight intervals less than 3 months apart, and those who died were reviewed to obtain a final weight that did not reflect a sudden decrease associated with the terminal event. Seven subjects were excluded because of gastric bypass surgery during the follow-up period. The final sample included 3834 subjects.
      Weight change was classified as those whose weight remained stable, those whose weight increased, and those who lost weight between the initial and the subsequent clinic visit. Delta weight was calculated and divided by the time between weights to provide a change per year. Similar calculations were performed to express weight change as a percentage of baseline weight. An alteration in weight (increase or decrease)1.0 lb was considered a change. The analyses were repeated using a2.0-lb weight change to determine whether a higher threshold had an appreciable effect on the results. Historical information that was recorded at the time of the exercise test included previous myocardial infarction by history or Q wave, cardiac procedures, heart failure, hypertension, hypercholesterolemia (>220 mg/dL), claudication, chronic obstructive pulmonary disease, cancer, renal disease, diabetes, stroke, smoking status (current, past), and use of cardiac medications.

      Exercise Testing

      Patients underwent symptom-limited treadmill testing using an individualized ramp treadmill protocol.
      • Myers J.
      • Buchanan N.
      • Smith D.
      • et al.
      Individualized ramp treadmill: observations on a new protocol.
      Standard criteria for termination were used, including moderately severe angina, >2.0 mm abnormal ST depression, a sustained decrease in systolic blood pressure, or serious rhythm disturbances. The Borg 6-20 perceived exertion scale was used to quantify degree of effort.
      • Borg G.A.V.
      Borg's Perceived Exertion and Pain Scales.
      Blood pressure was taken manually, and exercise capacity (in metabolic equivalents [METs]) was estimated from peak treadmill speed and grade.
      American College of Sports Medicine
      Guidelines for Exercise Testing and Prescription.
      No test was classified as indeterminate, medications were not withheld, and age-predicted maximal target heart rates were not used as end points. The exercise tests were performed, analyzed, and reported using a standard protocol incorporating a computerized database with all definitions and measurements prospectively defined.
      • Shue P.
      • Froelicher V.F.
      EXTRA: an expert system for exercise test reporting.

      Outcomes

      The main outcome variable was all-cause mortality; cardiovascular deaths and other specific causes of death also were recorded. The California Health Department Service and Social Security Death Indices were used to ascertain the vital status of each patient as of December 31, 2008. Accuracy of deaths was reviewed by 2 clinicians blinded to exercise test results and confirmed using the Veterans Affairs computerized medical records.

      Statistical Analysis

      NCSS software (Kayesville, Utah) was used for all statistical analyses. Unpaired t tests were used for comparisons of continuous variables, and chi-square tests were used to compare dichotomous variables between groups. Survival analysis was performed using Cox proportional hazards analysis to determine which clinical and exercise test variables were independently associated with all-cause and cardiovascular death. Survival analyses were adjusted for age, exercise capacity, and cardiovascular disease. Hazard ratios were calculated along with their 95% confidence intervals. The proportional hazards assumption was evaluated and confirmed using the scaled Schoenfeld residual. Kaplan–Meier survival curves were constructed between subjects exhibiting weight gain and weight loss, and compared using a log-rank test. A delta weight score was created so that units of weight change per year in pounds would be analogous to METs in a scale of 16, ranging from −8 to +8 lbs/year, allowing comparable change in percentage risk.

      Results

      The sample included 3834 men with a mean age of 59 years (standard deviation [SD]±11), 75% of whom were white, 10% were black, 8% were Hispanic, and 7% were other. The mean follow-up period was 6.8 years (SD±3.1), and there were 314 deaths (8.2%) for an annual mortality of 1.1%; 23% of the deaths were due to cardiovascular disease. Weight parameters are presented in Table 1. The obesity paradox was present as reported previously in the VETS cohort with the survivors weighing 201±38 lbs at baseline and those who died weighing 192±40 lbs (P<.001). The population on average gained 0.8 lbs (SD±20) or 0.19 lbs/year (SD±6); 52% of the men gained weight or stayed the same, and 48% lost weight. Those who lost weight were on average 5 years older than those who gained weight (P<.001). The survivors gained 0.39±6 lbs/year, and those who died lost 2.0±9 lbs/year (P<.001).
      Table 1Weight Parameter Comparisons Based on Vital Status and Weight Gain or Loss
      Weight Parameters (mean ± SD)Entire Population (n = 3834)Survivors (n = 3520)Deaths (All Cause) (n = 314)Deaths (Cardiac) (n = 72)Weight Gain (n = 1995)Weight Loss (n = 1839)
      P Value
      P value for comparison with survivors.
      P Value
      P value for comparison with survivors.
      P Value
      P value for comparison between weight gain and loss groups.
      Age, y59±1159±1167±11.00068±10.00057±1062±11.000
      Weight, lbs200±39201±38192±40.000188±37.008202±39198±38.004
      Height, in69±369±369±3.0668±4.00469±369±3.16
      BMI, kg/m229±529±528±5.00028±5.2229±529±5.02
      Obesity (%)1455 (38)1350 (38)105 (33).0927 (38).99783 (39)672 (37).09
      Delta y6.8±36.9±35.9±3.0005.9±3.036.8±36.8±3.79
      Delta weight0.81±201.62±20−8.30±23.000−6.1±16.00315±14−14.5±13.000
      Delta weight/y0.19±60.39±5.7−2.0±8.6.000−0.97±11.093.1±5.2−3.0±5.1.000
      Delta weight %0.52±9.80.94±9.54.1±11.000−3.1±8.0017.5±6.9−7.1±6.0.000
      Delta weight %/y0.12±30.22±2.8−1.0.±4.4.000−0.49±6.081.6±2.7−1.5±2.4.000
      BMI=body mass index; SD=standard deviation. Obesity=BMI>30%; Delta years=time between clinic weights; Delta weight=change from weight between visits; Delta weight %=percent change from baseline weight.
      low asterisk P value for comparison with survivors.
      P value for comparison between weight gain and loss groups.
      Differences in body dimensions, historical data, risk factors, and exercise test responses between those who died and those who survived in the weight loss and weight gain groups are presented in Table 2. In the weight loss group, those who died had greater weight loss and weight loss per year than those who survived. Although the prevalence of pulmonary disease, renal failure, and cancer did not differ between the groups, both coronary artery disease and heart failure were more prevalent in those who died. Exercise capacity also was significantly lower in those who died (8.3±3.0 vs 6.1±3.0 METs, P<.01). Among those who gained weight, no differences were observed between those who died and those who survived in terms of weight parameters, but cardiovascular disease was significantly more prevalent among those who died. Similar to the weight loss group, exercise capacity was 33% higher among those who survived.
      Table 2Clinical, Historical, and Exercise Test Differences Between Those Who Survived and Those Who Died in the Group that Lost Weight (Mean±SD or Number [%])
      VariablesWeight Loss (N = 1839)P ValueWeight Gain (N = 1995)P Value
      Died (n = 211)Survived (n = 1628)Died (n = 103)Survived (n = 1892)
      Age, y69±1061±11<.0162±1057±10<.01
      Weight Parameters
       Weight, lbs188±37199±38<.01199±45202±38.54
       Height, in69±369±3.1369±369±3.40
       BMI, kg/m228±529±5<.0129±529±5.63
       Delta weight−19.7±16.3−13.9±12.6<.0115.0±16.215.0±13.8.98
       Delta weight/y−4.9±8−2.8±4.5<.013.8±6.83.1±5.2.17
       Delta wt %−9.9±7.5−6.7±5.8<.017.7±8.17.5±6.9.82
       Delta wt %/y−2.5±3.9−1.4±2.2<.012.0±3.81.6±2.6.08
       Obesity (%) [>30% BMI]59 (28)618 (38).00546 (45)737 (39).24
      Risk factors (%)
       Diabetes47 (22)343 (21).6826 (25)328 (17).04
       Hypertension130 (62)955 (59).4159 (57)1022 (54).42
       Hypercholesterolemia56 (27)736 (45)<.0139 (38)809 (43).96
       Family CAD44 (21)392 (24).2927 (26)513 (27).04
       Ever smoked146 (69)1070 (66).3272 (70)1279 (68).23
       Current smoking49 (23)361 (22).7329 (28)438 (23).24
       Alcohol abuse (history)4 (2)106 (7)<.016 (6)138 (7).57
       Medical history (%)
       Pulmonary disease23 (11)147 (9).3714 (14)159 (8).06
       Chronic renal insufficiency2 (0.9)24 (1.5).541 (1)19 (1).97
       Cancer4 (1.9)28 (1.7).85034 (2).17
      Cardiovascular diseases (%)
       Typical angina20 (9)114 (7).1915 (15)147 (8).01
       Myocardial infarction54 (26)220 (13)<.0129 (28)239 (13)<.01
       CABG46 (22)96 (6)<.0117 (16)95 (5)<.01
       PCI17 (8)84 (5).089 (9)89 (5).06
       Any CAD87 (41)402 (25)<.0147 (46)440 (23)<.01
       Stroke8 (4)51 (3).616 (6)40 (2).01
       Claudication25 (12)53 (3)<.014 (4)56 (3).59
       VHD7 (3)26 (1.6).07023 (1).26
       Heart failure38 (18)70 (4)<.0116 (15)88 (5)<.01
       Any HD93 (44)403 (25)<.0157 (55)529 (28)<.01
      Exercise responses
       Exercise capacity (METs)6.1±38.3±3.0<.016.7±2.38.9±3.1<.01
       Peak heart rate128±23139±22<.01129±23142±22<.01
       Peak perceived exertion17±217±2.1917±217±2.68
       ST depression1.0 mm44 (21)269 (17).1127 (26)282 (15).002
      BMI=body mass index; CABG=coronary artery bypass graft; CAD=coronary artery disease; HD=heart disease; MET=metabolic equivalent; PCI=percutaneous coronary intervention; VHD=valvular heart disease. P values represent comparisons within weight gain and weight loss groups. Any coronary artery disease=coronary artery bypass graft, myocardial infarction, percutaneous coronary intervention, or typical angina. Any heart disease=any of the conditions above.
      The annual mortality for those who gained weight was 0.8% compared with 1.7% for those who lost weight (P<.001, Figure 1). Stepwise age-adjusted proportional hazards analyses were performed initially for weight at the time of treadmill testing and then sequentially adding delta weight per year and exercise capacity (Table 3). Both exercise capacity and delta weight per year were significantly and independently associated with time to death, with the risk decreasing 18% for each MET achieved and increasing 4% per pound lost per year.
      Figure thumbnail gr1
      Figure 1Kaplan-Meier survival curves for subjects exhibiting weight gain and weight loss. Log rank test P < .01 between groups.
      Table 3Age-Adjusted Proportional Hazards Analysis with All-Cause Mortality as the Outcome for Each Predictor Alone and Combined
      Hazard Ratio95% Confidence IntervalP ValueRegression Coefficient
      Univariate Analysis
       Baseline weight1.000.99-1.00.98−0.00
       Exercise capacity0.820.78-0.86<.0001−0.20
       Delta lbs/y0.960.94-0.97<.0001−0.04
      Multivariate Analysis
      Model 1
       Baseline weight0.990.99-1.00.86−0.00
       Delta lbs/y0.960.95-0.97<.0001−0.04
      Model 2
       Exercise capacity0.820.78-0.86<.0001−0.19
       Delta lbs/y0.960.95-0.97<.0001−0.03
      Model 3
       Baseline weight0.990.99-1.00.13−0.002
       Exercise capacity0.820.78-0.86<.0001−0.20
       Delta lbs/y0.960.95-0.97<.0001−0.04
      Table 4 shows causes of death for the entire cohort, for those who lost weight, and for those who gained weight. Subjects who lost weight had a higher proportion of deaths attributable to heart failure and cancer, whereas deaths due to gastrointestinal causes were higher in the weight gain group. By including cancer, heart failure, renal failure, and 7 additional subjects who died of human immunodeficiency virus/autoimmune deficiency syndrome or hepatitis/cirrhosis, a total of 58.4% of the subjects in the weight loss group died of causes associated with cachexia (vs 26.8% in the weight gain group). The age, exercise capacity, and cardiovascular disease-adjusted hazard ratios for all-cause and cardiovascular mortality are presented as relative risks for subjects who exhibited no change in weight, weight gain, and weight loss in Table 5, with no change in weight as the reference group. The weight gain groups had 36% and 24% reductions in all-cause and cardiovascular mortality, respectively, versus the no change group. In contrast, the weight loss groups demonstrated 49% and 25% higher all-cause and cardiovascular mortality, respectively, versus the no change groups. These differences were significant for all-cause (P<.001) but not cardiovascular mortality (P=.37 and .31 for weight loss and weight gain groups, respectively). The results were similar when the analyses were repeated using a 2-lb weight change as the criterion for loss or gain.
      Table 4Causes of Death in the Entire Cohort and the Weight Loss and Weight Gain Groups
      All Deaths (n = 314)Deaths in Weight Loss Group (n = 206)Deaths in Weight Gain/No Change Group (n = 108)P Value
      Comparison between weight loss and weight gain groups.
      Causes of Death (%)
       Heart failure25 (7.9)21 (10.2)4 (3.7).01
       Cancer93 (29.6)78 (37.9)15 (13.9).000
       Gastrointestinal44 (14.0)17 (8.3)27 (25.0).000
       Pulmonary42 (13.4)25 (12.1)17 (15.7).37
       All cardiac72 (22.9)47 (22.8)25 (23.1).95
       Renal24 (7.6)14 (6.8)10 (9.3).43
       Neurologic25 (7.8)16 (7.8)9 (8.3).86
       Accidents/suicide1 (0.3)01 (0.9).16
       Others13 (4.1)9 (4.4)4 (3.7).78
      low asterisk Comparison between weight loss and weight gain groups.
      Table 5Relative Risks for All-Cause and Cardiovascular Mortality Among Weight Gain and Weight Loss Groups Relative to the No Change (Reference) Group, with Multivariate Adjustments
      AdjustmentReferenceWeight GainWeight Loss
      All-Cause Mortality
       Age1.00.61 (0.47-0.78)1.60 (1.26-2.03)
       Age and exercise capacity1.00.63 (0.49-0.81)1.51 (1.19-1.92)
       Age, exercise capacity, and cardiovascular disease1.00.64 (0.50-0.83)1.49 (1.17-1.89)
      Cardiovascular Mortality
       Age1.00.66 (0.39-1.11)1.46 (0.89-2.41)
       Age and exercise capacity1.00.71 (0.42-1.18)1.35 (0.82-2.23)
       Age, exercise capacity, and cardiovascular disease1.00.76 (0.45-1.28)1.25 (0.76-2.07)
      Cardiovascular disease includes history of myocardial infarction, coronary artery disease, heart failure, stroke, or bypass surgery; 95% confidence intervals are in parentheses.

      Discussion

      We previously observed the existence of an “obesity paradox” in veterans referred for exercise testing for clinical reasons, in which overweight individuals exhibited better survival than normal weight subjects.
      • McAuley P.
      • Myers J.
      • Abella J.
      • Froelicher V.
      Body mass, fitness and survival in veteran patients: another obesity paradox?.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      • Lissin L.W.
      • Gauri A.J.
      • Froelicher V.F.
      • Ghayoumi A.
      • Myers J.
      • Giacommini J.
      The prognostic value of body mass index and standard exercise testing in male veterans with congestive heart failure.
      Most previous studies addressing the obesity paradox have been limited to a measure of body dimensions at a single baseline visit, and the effects of changes in weight have not been fully explored. The current results suggest that this phenomenon is explained in part by weight loss associated with conditions linked to muscle wasting. Weight loss had a significant and independent relationship to all-cause mortality, as demonstrated by a 4% increase in mortality per pound lost over the mean follow-up of 7 years, in contrast with the comparatively lower mortality rate among those exhibiting weight gain (Figure 1).
      One possible explanation for the weight loss and higher mortality we observed is occult disease. It is surprising that clinical assessment of baseline health status did not detect the illnesses that cause weight loss and mortality, but the current data suggest that these illnesses exist. The decline in weight was evident even after accounting for weight loss associated with terminal events. The concept that the obesity paradox in our sample is partly attributable to occult disease is underscored by the fact that two thirds of the deaths occurred in the weight loss group, and approximately 60% of these were due to conditions associated with muscle wasting. Notably, most of these conditions were not apparent at the time of the treadmill test but were diagnoses that were made during the mean 7-year follow-up. Previous studies on the obesity paradox have generally reported all-cause mortality, but few have assessed specific causes of death. The majority of these studies have removed subjects with potentially wasting conditions at baseline, as was the case in our previous studies;
      • McAuley P.
      • Myers J.
      • Abella J.
      • Froelicher V.
      Body mass, fitness and survival in veteran patients: another obesity paradox?.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      • McAuley P.
      • Pittsley J.
      • Myers J.
      • Abella J.
      • Froelicher V.F.
      Fitness and fatness as mortality predictors in healthy older men: the Veterans Exercise Testing Study.
      therefore, little is known regarding the effect of these conditions on the obesity paradox. In addition to serial measures of weight status, a unique feature of the current study was the fact that causes of death were carefully determined from computerized medical records during the follow-up. The higher mortality associated with weight loss and chronic conditions, including cancer and heart failure (Table 4), suggests that occult disease contributed in part to the obesity paradox.
      Although the obesity paradox has been described in a number of clinical populations over the last decade,
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Arena R.
      • Lavie C.J.
      The obesity paradox and outcome in heart failure: is excess bodyweight truly protective?.
      there are few studies assessing weight change.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Myrskyla M.
      • Chang V.W.
      Weight change, initial BMI, and mortality among middle- and older-aged adults.
      • Østergaard J.N.
      • Grønbaek M.
      • Schnohr P.
      • Sørensen T.I.
      • Heitmann B.L.
      Combined effects of weight loss and physical activity on all-cause mortality of overweight men and women.
      Even in populations in whom excess weight has been associated with poor health outcomes, weight loss in some studies has been associated with increased mortality.
      • Strandberg T.E.
      • Strandberg A.Y.
      • Salomaa V.V.
      • et al.
      Explaining the obesity paradox: cardiovascular risk, weight change, and mortality during long-term follow-up in men.
      • Lavie C.J.
      • Milani R.V.
      • Artham S.M.
      • Patel D.A.
      • Ventura H.O.
      The obesity paradox, weight loss, and coronary disease.
      • Blair S.N.
      • Shaten J.
      • Brownell K.
      • Collins G.
      • Lissner L.
      Body weight change, all-cause mortality, and cause-specific mortality in the Multiple Risk Factor Intervention Trial.
      In the Multiple Risk Factor Intervention Trial, weight variation (both gain and loss) over a 6- to 7-year period was associated with higher cardiovascular and all-cause mortality compared with those whose weight remained stable.
      • Blair S.N.
      • Shaten J.
      • Brownell K.
      • Collins G.
      • Lissner L.
      Body weight change, all-cause mortality, and cause-specific mortality in the Multiple Risk Factor Intervention Trial.
      Those who lost weight had slightly worse outcomes than those who gained weight. In the Established Populations for Epidemiologic Studies of the Elderly Study, those who lost 10% or more of body weight between age 50 years and old age had the highest risk of mortality (relative risk=1.6). Exclusion of participants who lost 10% or more of their weight and adjustment for health status eliminated the higher risk of death associated with low weight. The authors suggested that the inverse association between weight and mortality in old age reflects illness-related weight loss from heavier weight in middle age.
      • Losonczy K.G.
      • Harris T.B.
      • Cornoni-Huntley J.
      • et al.
      Does weight loss from middle age to old age explain the inverse weight mortality relation in old age?.
      Strandberg and associates
      • Strandberg T.E.
      • Strandberg A.Y.
      • Salomaa V.V.
      • et al.
      Explaining the obesity paradox: cardiovascular risk, weight change, and mortality during long-term follow-up in men.
      evaluated 1114 men in 1974 and again in 2000, and reported that among groups who gained, lost, or maintained body weight, the group who lost weight had the highest rate of late-life comorbidities and the greatest total mortality, with a relative risk of 2.0 compared to the group with stable weight across the follow-up. Most recently, Lavie et al
      • Lavie C.J.
      • Milani R.V.
      • Artham S.M.
      • Patel D.A.
      • Ventura H.O.
      The obesity paradox, weight loss, and coronary disease.
      followed 529 cardiac rehabilitation participants for 3 years and observed only a slight trend for lower mortality among patients who lost weight.
      The concept that the relationship between lower weight and higher mortality in the current study and others
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Arena R.
      • Lavie C.J.
      The obesity paradox and outcome in heart failure: is excess bodyweight truly protective?.
      • Oreopoulos A.
      • Padwal R.
      • Kalantar-Zadeh K.
      • Fonarow G.C.
      • Norris C.M.
      • McAlister F.A.
      Body mass index and mortality in heart failure: a meta-analysis.
      • McAuley P.
      • Myers J.
      • Abella J.
      • Froelicher V.
      Body mass, fitness and survival in veteran patients: another obesity paradox?.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      • McAuley P.
      • Pittsley J.
      • Myers J.
      • Abella J.
      • Froelicher V.F.
      Fitness and fatness as mortality predictors in healthy older men: the Veterans Exercise Testing Study.
      • Sorensen T.I.A.
      Weight loss causes increased mortality: pros.
      • Ingram D.D.
      • Mussolino M.E.
      Weight loss from maximum body weight and mortality: the Third National Health and Nutrition Examination Survey Linked Mortality File.
      • Lissin L.W.
      • Gauri A.J.
      • Froelicher V.F.
      • Ghayoumi A.
      • Myers J.
      • Giacommini J.
      The prognostic value of body mass index and standard exercise testing in male veterans with congestive heart failure.
      • Strandberg T.E.
      • Strandberg A.Y.
      • Salomaa V.V.
      • et al.
      Explaining the obesity paradox: cardiovascular risk, weight change, and mortality during long-term follow-up in men.
      • Lavie C.J.
      • Milani R.V.
      • Artham S.M.
      • Patel D.A.
      • Ventura H.O.
      The obesity paradox, weight loss, and coronary disease.
      • Blair S.N.
      • Shaten J.
      • Brownell K.
      • Collins G.
      • Lissner L.
      Body weight change, all-cause mortality, and cause-specific mortality in the Multiple Risk Factor Intervention Trial.
      • Losonczy K.G.
      • Harris T.B.
      • Cornoni-Huntley J.
      • et al.
      Does weight loss from middle age to old age explain the inverse weight mortality relation in old age?.
      • Morse S.A.
      • Gulati R.
      • Reisin E.
      The obesity paradox and cardiovascular disease.
      results from unintentional weight loss is important to consider. Numerous studies in cardiac rehabilitation and other settings in which clinically supervised (intentional) weight loss occurred have shown extensive health benefits, including a lower incidence of cardiovascular events and better overall survival.
      • Poirier P.
      • Giles T.D.
      • Bray G.A.
      • et al.
      Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss.
      • Lavie C.J.
      • Milani R.V.
      • Ventura H.O.
      Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
      • Klein S.
      • Burke L.E.
      • Bray G.A.
      • et al.
      American Heart Association Council on Nutrition, Physical Activity, and Metabolism Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism.
      • Savage P.D.
      • Brochu M.
      • Poehlman E.T.
      • Ades P.A.
      Reduction in obesity and coronary risk factors after high calorie exercise training in overweight coronary patients.
      • Ades P.A.
      • Savage P.D.
      • Harvey-Berino J.
      The treatment of obesity in cardiac rehabilitation.
      Intentional weight loss has been associated with marked reductions in the metabolic syndrome, inflammatory markers, lipids, prevalence of hypertension, and better glucose tolerance.
      • Horton E.S.
      Effects of lifestyle changes to reduce the risks of diabetes and associated cardiovascular risks: results from large scale efficacy trials.
      Several of the original observations on the obesity paradox were made in the context of patients with heart failure,
      • Lissin L.W.
      • Gauri A.J.
      • Froelicher V.F.
      • Ghayoumi A.
      • Myers J.
      • Giacommini J.
      The prognostic value of body mass index and standard exercise testing in male veterans with congestive heart failure.
      • Horwich T.B.
      • Fonarow G.C.
      • Hamilton M.A.
      • MacLellan W.R.
      • Woo M.A.
      • Tillisch J.H.
      The relationship between obesity and mortality in patients with heart failure.
      and studies have shown that weight loss can reduce left ventricular mass and ameliorate both systolic and diastolic dysfunction.
      • Dayl S.U.
      • Kasikcioglu H.
      • Uslu N.
      • et al.
      Influence of weight loss on myocardial performance index.
      • Alpert M.A.
      • Fraley M.A.
      • Birchem J.A.
      • Senkottaiyan N.
      Management of obesity cardiomyopathy.
      Although long-term studies are lacking, gastric bypass surgery has been consistently associated with considerable reductions in markers of cardiovascular risk,
      • Batsis J.A.
      • Sarr M.G.
      • Collazo-Clavell M.L.
      • et al.
      Cardiovascular risk after bariatric surgery for obesity.
      • Leichman J.G.
      • Wilson E.B.
      • Scarborough T.
      • et al.
      Dramatic reversal of derangements in muscle metabolism and left ventricular function after bariatric surgery.
      and limited available data suggest that surgical weight loss results in reduced morbidity and mortality.
      • Picot J.
      • Jones J.
      • Colquitt J.L.
      • et al.
      The clinical effectiveness and cost-effectiveness of bariatric (weight loss) surgery for obesity: asystematic review and economic evaluation.
      Despite the obesity paradox, these studies support clinically supervised weight reduction, particularly in patients with or at high risk for cardiovascular disease.

      Limitations

      A complete explanation of the contribution of weight change to the obesity paradox would require a detailed examination of latent disease, which would be a difficult undertaking. Our findings are limited to men. The VETS study consists of men referred for exercise testing for clinical reasons, and although cardiovascular and other diseases were ruled out in the majority, the test was conducted for clinical reasons. We do not have data on waist circumference or fat distribution, factors that have been suggested to contribute the obesity paradox;
      • Beddhu S.
      • Pappas L.M.
      • Ramkumar N.
      • Samore M.
      Effects of body size and body composition on survival in hemodialysis patients.
      • Tanko L.B.
      • Christiansen C.
      Can the obesity paradox be explained by the protective effects of peripheral adiposity?.
      we only measured body mass index, which may not adequately capture the health effects of obesity.
      • Janssen I.
      • Katzmarzyk P.T.
      • Ross R.
      Waist circumference and not body mass index explains the obesity paradox.
      Finally, we have previously reported that the obesity paradox is negated in subjects with high fitness,
      • McAuley P.
      • Myers J.
      • Abella J.
      • Froelicher V.
      Body mass, fitness and survival in veteran patients: another obesity paradox?.
      • McAuley P.A.
      • Kokkinos P.F.
      • Oliveira R.B.
      • Emerson B.T.
      • Myers J.N.
      Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
      and we do not have serial measures of fitness in our subjects. Therefore, the extent to which fitness level interacted with weight change and mortality is unknown.

      Conclusions

      Weight loss over a mean 7-year period was related to higher all-cause mortality in veterans who were referred for exercise testing for clinical reasons. The obesity paradox in our sample seems to be explained in part by non-volitional weight loss related to latent disease that contributes to early mortality. The obesity paradox is a complex phenomenon that requires additional study, and future studies should consider weight change when evaluating the longitudinal association among health, overweight/obesity, and outcomes.

      References

        • Poirier P.
        • Giles T.D.
        • Bray G.A.
        • et al.
        Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss.
        Arterioscler Thromb Vasc Biol. 2006; 26: 968-976
        • Lavie C.J.
        • Milani R.V.
        • Ventura H.O.
        Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss.
        J Am Coll Cardiol. 2009; 53: 1925-1932
        • Klein S.
        • Burke L.E.
        • Bray G.A.
        • et al.
        American Heart Association Council on Nutrition, Physical Activity, and Metabolism.
        Circulation. 2004; 110: 2952-2967
        • Arena R.
        • Lavie C.J.
        The obesity paradox and outcome in heart failure: is excess bodyweight truly protective?.
        Future Cardiol. 2010; 6: 1-6
        • Oreopoulos A.
        • Padwal R.
        • Kalantar-Zadeh K.
        • Fonarow G.C.
        • Norris C.M.
        • McAlister F.A.
        Body mass index and mortality in heart failure: a meta-analysis.
        Am Heart J. 2008; 156: 13-22
        • McAuley P.
        • Myers J.
        • Abella J.
        • Froelicher V.
        Body mass, fitness and survival in veteran patients: another obesity paradox?.
        Am J Med. 2007; 120: 518-524
        • McAuley P.A.
        • Kokkinos P.F.
        • Oliveira R.B.
        • Emerson B.T.
        • Myers J.N.
        Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years.
        Mayo Clinic Proc. 2010; 85: 115-121
        • McAuley P.
        • Pittsley J.
        • Myers J.
        • Abella J.
        • Froelicher V.F.
        Fitness and fatness as mortality predictors in healthy older men: the Veterans Exercise Testing Study.
        J Gerontol A Biol Sci Med Sci. 2009; 64: 695-699
        • Klein S.
        • Allison D.B.
        • Heymsfield S.B.
        • et al.
        Waist Circumference and Cardiometabolic Risk: a Consensus Statement from Shaping America's Health: Association for Weight Management and Obesity Prevention; NAASO, the Obesity Society; the American Society for Nutrition; and the American Diabetes Association.
        Obesity. 2007; 15: 1061-1067
        • Lee I.M.
        • Paffenbarger Jr, R.S.
        Is weight loss hazardous?.
        Nutr Rev. 1996; 54: S116-S124
        • Sorensen T.I.A.
        Weight loss causes increased mortality: pros.
        Obes Rev. 2003; 4: 3-7
        • Yang D.
        • Fontaine K.R.
        • Wang C.
        • Allison D.B.
        Weight loss causes increased mortality: cons.
        Obes Rev. 2003; 4: 9-16
        • Myrskyla M.
        • Chang V.W.
        Weight change, initial BMI, and mortality among middle- and older-aged adults.
        Epidemiology. 2009; 20: 840-848
        • Østergaard J.N.
        • Grønbaek M.
        • Schnohr P.
        • Sørensen T.I.
        • Heitmann B.L.
        Combined effects of weight loss and physical activity on all-cause mortality of overweight men and women.
        Int J Obes. 2010; 34: 760-769
        • Ingram D.D.
        • Mussolino M.E.
        Weight loss from maximum body weight and mortality: the Third National Health and Nutrition Examination Survey Linked Mortality File.
        Int J Obes. 2010; 34: 1044-1050
        • Ross R.
        • Janiszewski P.M.
        Is weight loss the optimal target for obesity-related cardiovascular disease risk reduction?.
        Can J Cardiol. 2008; 24: 25D-31D
        • Lissin L.W.
        • Gauri A.J.
        • Froelicher V.F.
        • Ghayoumi A.
        • Myers J.
        • Giacommini J.
        The prognostic value of body mass index and standard exercise testing in male veterans with congestive heart failure.
        J Card Fail. 2002; 8: 206-215
        • Myers J.
        • Buchanan N.
        • Smith D.
        • et al.
        Individualized ramp treadmill: observations on a new protocol.
        Chest. 1992; 101: 236S-241S
        • Borg G.A.V.
        Borg's Perceived Exertion and Pain Scales.
        Human Kinetics, Champaign, IL1998
        • American College of Sports Medicine
        Guidelines for Exercise Testing and Prescription.
        8th ed. Lippincott, Williams, & Wilkins, Baltimore2010
        • Shue P.
        • Froelicher V.F.
        EXTRA: an expert system for exercise test reporting.
        J Non-Invasive Testing. 1998; II-4: 21-27
        • Strandberg T.E.
        • Strandberg A.Y.
        • Salomaa V.V.
        • et al.
        Explaining the obesity paradox: cardiovascular risk, weight change, and mortality during long-term follow-up in men.
        Eur Heart J. 2009; 30: 1720-1727
        • Lavie C.J.
        • Milani R.V.
        • Artham S.M.
        • Patel D.A.
        • Ventura H.O.
        The obesity paradox, weight loss, and coronary disease.
        Am J Med. 2009; 122: 1106-1114
        • Blair S.N.
        • Shaten J.
        • Brownell K.
        • Collins G.
        • Lissner L.
        Body weight change, all-cause mortality, and cause-specific mortality in the Multiple Risk Factor Intervention Trial.
        Ann Intern Med. 1993; 119: 749-757
        • Losonczy K.G.
        • Harris T.B.
        • Cornoni-Huntley J.
        • et al.
        Does weight loss from middle age to old age explain the inverse weight mortality relation in old age?.
        Am J Epidemiol. 1995; 141: 312-321
        • Morse S.A.
        • Gulati R.
        • Reisin E.
        The obesity paradox and cardiovascular disease.
        Curr Hypertens Rep. 2010; 12: 120-126
        • Savage P.D.
        • Brochu M.
        • Poehlman E.T.
        • Ades P.A.
        Reduction in obesity and coronary risk factors after high calorie exercise training in overweight coronary patients.
        Am Heart J. 2003; 146: 317-323
        • Ades P.A.
        • Savage P.D.
        • Harvey-Berino J.
        The treatment of obesity in cardiac rehabilitation.
        J Cardiopulm Rehabil Prev. 2010; 30: 289-298
        • Horton E.S.
        Effects of lifestyle changes to reduce the risks of diabetes and associated cardiovascular risks: results from large scale efficacy trials.
        Obesity. 2009; 17: S43-S48
        • Horwich T.B.
        • Fonarow G.C.
        • Hamilton M.A.
        • MacLellan W.R.
        • Woo M.A.
        • Tillisch J.H.
        The relationship between obesity and mortality in patients with heart failure.
        J Am Coll Cardiol. 2001; 38: 789-795
        • Dayl S.U.
        • Kasikcioglu H.
        • Uslu N.
        • et al.
        Influence of weight loss on myocardial performance index.
        Heart Vessels. 2006; 21: 84-88
        • Alpert M.A.
        • Fraley M.A.
        • Birchem J.A.
        • Senkottaiyan N.
        Management of obesity cardiomyopathy.
        Exp Rev Cardiovasc Ther. 2005; 3: 225-230
        • Batsis J.A.
        • Sarr M.G.
        • Collazo-Clavell M.L.
        • et al.
        Cardiovascular risk after bariatric surgery for obesity.
        Am J Cardiol. 2008; 102: 930-937
        • Leichman J.G.
        • Wilson E.B.
        • Scarborough T.
        • et al.
        Dramatic reversal of derangements in muscle metabolism and left ventricular function after bariatric surgery.
        Am J Med. 2008; 121: 966-973
        • Picot J.
        • Jones J.
        • Colquitt J.L.
        • et al.
        The clinical effectiveness and cost-effectiveness of bariatric (weight loss) surgery for obesity: asystematic review and economic evaluation.
        Health Technol Assess. 2009; 13 (215-357): 1-190
        • Beddhu S.
        • Pappas L.M.
        • Ramkumar N.
        • Samore M.
        Effects of body size and body composition on survival in hemodialysis patients.
        J Am Soc Nephrol. 2003; 14: 2366-2372
        • Tanko L.B.
        • Christiansen C.
        Can the obesity paradox be explained by the protective effects of peripheral adiposity?.
        Arch Intern Med. 2005; 165: 1796-1797
        • Janssen I.
        • Katzmarzyk P.T.
        • Ross R.
        Waist circumference and not body mass index explains the obesity paradox.
        Am J Clin Nutr. 2004; 79: 379-384