Obesity and mortality in older adults

The most commonly used measure of overweight and obesity is BMI, defined by WHO as cut-off points of ⩾25 and ⩾30 kg/m², respectively^{( 21 )}. However, the relationship between BMI and mortality in older age is controversial. A systematic review of older adults aged ⩾65 years found that BMI in the overweight range is not associated with a significantly increased mortality risk, and BMI in the moderately obese range is only associated with a modest increase in mortality risk⁽ Reference Janssen and Mark ^{4 )}. Similarly, a more recent large meta-analysis of nearly 200 000 individuals aged 65 or older showed a U-shaped relationship between BMI and mortality, with the lowest risk seen in those with a BMI between 24·0 and 30·0 kg/m² and risk only began to increase when BMI exceeded 33 kg/m^{2 (} Reference Winter, MacInnis and Wattanapenpaiboon ^{5 )}. Ageing is associated with significant changes in body composition with a substantial reduction in fat-free mass and muscle mass and an increase in visceral fat, even if the body weight remains unchanged⁽ Reference Zamboni, Mazzali and Zoico ^{3 ,} Reference Baumgartner ^{14 ,} Reference Zamboni, Mazzali and Fantin ^{15 )}. Thus BMI depends not only on adiposity but also on the loss of muscle mass which has opposing effects on mortality⁽ Reference Allison, Zhu and Plankey ^{13 )}, so that BMI may not be a good indicator of obesity in the elderly. It is suggested that measures of adiposity such as waist circumference (WC) or the waist:hip ratio, which better reflect visceral fat, may be more useful at assessing obesity risk as they are better at predicting CVD and mortality in older subjects⁽ Reference Czernichow, Kengne and Stamatakis ^{22 –} Reference de Hollander, Bemelmans and Boshuizen ^{26 )}. Central obesity has been defined as a WC > 102 cm for men and >88 cm for women, or a waist:hip ratio ⩾0·9 for men and ⩾0·85 for women^{( 27 )}.

Sarcopenia

Sarcopenia, the age-related loss of muscle mass and decline in muscle strength, is strongly associated with physical disability, poor quality of life and frailty⁽ Reference Baumgartner, Koehler and Gallagher ^{28 )}. Several different definitions of sarcopenia have been used in the literature but to date no consensus definition has been universally adopted⁽ Reference Cruz-Jentoft, Baeyens and Bauer ^{29 )}. Measurement methods such as computerised tomography, dual energy X-ray absorptiometry (DXA) and bioelectrical impedance analysis (BIA) have been used to assess total or skeletal muscle mass⁽ Reference Zamboni, Mazzali and Fantin ^{15 )}. Sarcopenia was originally defined on the basis of appendicular skeletal muscle mass, assessed by DXA and adjusted for height. Those whose height-adjusted appendicular skeletal muscle mass were two standard deviations below the reference for healthy younger persons were considered to be sarcopenic⁽ Reference Baumgartner, Koehler and Gallagher ^{28 )}. Subsequently, the definition of sarcopenia has evolved from a focus on muscle mass to muscle strength and physical function. Recently, a definition of sarcopenia has been suggested by The European Working Group on Sarcopenia in Older People which proposed the presence of both low muscle mass and low muscle function (strength or performance)⁽ Reference Cruz-Jentoft, Baeyens and Bauer ^{29 )}.

Sarcopenia and mortality and cardiovascular risk

Prospective studies have shown consistent associations between low muscle mass (as measured by DXA), BIA and mid-arm muscle circumference) or muscle strength as measured by hand grip and an increased risk of mortality⁽ Reference Wijnhoven, Snijder and van Bokhorst-de van der Schueren ^{30 –} Reference Arango-Lopera, Arroyo and Gutiérrez-Robledo ^{37 )}. Some studies suggest that decline in muscle strength is a stronger independent risk factor for mortality compared with muscle mass⁽ Reference Kim, Lim and Choi ^{38 ,} Reference Filippin, Teixeira and da Silva ^{39 )}.

Several mechanisms underlying age-related muscle loss have been recognised, including neuronal and hormonal changes, poor nutrition, physical inactivity and inflammation⁽ Reference Zamboni, Mazzali and Fantin ^{15 ,} Reference Jensen ^{40 )}. Thus sarcopenia shares many pathological mechanisms with atherosclerosis, including insulin resistance and inflammation⁽ Reference Zamboni, Mazzali and Fantin ^{15 )}. Population studies have shown sarcopenia to be associated with metabolic impairments, including insulin resistance⁽ Reference Zamboni, Mazzali and Fantin ^{15 ,} Reference Baumgartner, Koehler and Gallagher ^{28 ,} Reference Filippin, Teixeira and da Silva ^{39 )}. Low muscle mass has also been associated with CVD risk factors, including arterial stiffness and blood pressure⁽ Reference Han, Park and Kwon ^{41 ,} Reference Snijder, Henry and Visser ^{42 )}. However, the association with CVD disease is uncertain and few studies have examined the association between sarcopenia and CVD ⁽ Reference Atkins, Whincup and Morris ^{43 ,} Reference Stephen and Janssen ^{44 )}. In those that have there is suggestion that sarcopenia is associated with increased risk of CVD mortality⁽ Reference Atkins, Whincup and Morris ^{43 )}.

Sarcopenic obesity

Given the age-related changes in body composition, sarcopenia often coexists with an increase in fat mass (Fig. 1). Visceral fat and muscle mass are known to be interrelated from a pathogenic point of view and are reported to share common inflammatory pathways⁽ Reference Zamboni, Mazzali and Fantin ^{15 )}. In addition, sarcopenia reduces physical activity which results in decreased energy expenditure leading to increased risk of obesity⁽ Reference Zamboni, Mazzali and Fantin ^{15 )}. Alternatively, visceral fat induces inflammation which may contribute to the development of sarcopenia⁽ Reference Gregor and Hotamisligil ^{45 )}. The interplay between sarcopenia and rising trends in obesity in an ageing population is emerging as an important public health concern in the elderly. The term sarcopenic obesity was first introduced by Baumgartner⁽ Reference Baumgartner ^{14 )} and is defined by the combination of sarcopenia and obesity. Both obesity and sarcopenia are associated with metabolic disorders and are important causes of disability, morbidity and mortality⁽ Reference Zamboni, Mazzali and Fantin ^{15 )}. Therefore it is hypothesised that obesity and sarcopenia may act synergistically; sarcopenic obesity may have a greater effect on metabolic disorders, CVD and mortality than either obesity or sarcopenia alone. However, there are limited studies which have investigated the joint effect of sarcopenia and obesity on CVD and mortality, the majority of which have been cross-sectional studies or have focused on disability or physical function as outcomes⁽ Reference Zamboni, Mazzali and Fantin ^{15 ,} Reference Waters and Baumgartner ^{19 ,} Reference Prado, Wells and Smith ^{20 ,} Reference Dominguez and Barbagallo ^{46 )}.

Fig. 1.

(Colour online) Sarcopenic obesity with ageing. The interplay between sarcopenia and obesity. Adapted from Zamboni et al. ⁽ Reference Zamboni, Mazzali and Fantin ^{15 )}.

Sarcopenic obesity and mortality

Only a few population prospective studies have examined the association between sarcopenic obesity and the risk of all-cause mortality and there is suggestion that sarcopenic obese adults have the highest mortality risk. In an earlier report from the British Regional Heart study involving over 4000 men, aged 60–79 years followed up for 6 years, men with high WC (>102 cm) and low mid-arm muscle circumference (sarcopenic obese) showed a 55 % increase in mortality risk compared with non-sarcopenic, non-obese individuals⁽ Reference Wannamethee, Shaper and Lennon ^{33 )}. Similar findings were observed with extended 11-year follow-up with the sarcopenic obese group showing the highest mortality risk⁽ Reference Atkins, Whincup and Morris ^{43 )} (Fig. 2). A 14-year prospective study of participants from the National Health and Nutrition Examination Survey III found a similar association in sarcopenic obese women (based on skeletal muscle mass and body fat measurement from BIA), with a 29 % increase in mortality risk compared with those without sarcopenia or obesity⁽ Reference Batsis, Mackenzie and Barre ^{47 )}. However, no significant association was seen in this cohort between sarcopenic obesity and mortality in men. In the InCHIANTI study of 934 male and female participants aged 65 years and over followed for 6 years, no significant difference in mortality risk was reported across six sarcopenic obesity groups (defined using calf skeletal muscle and BMI) although sarcopenic obese adults showed the lowest survival⁽ Reference Cesari, Pahor and Lauretani ^{35 )}. A study which defined sarcopenic obesity using a measure of muscle strength instead of muscle mass showed a significant association between sarcopenic obesity and risk of mortality; adult men who were overweight (BMI ⩾ 25) and in the lowest grip strength tertile had the highest mortality risk⁽ Reference Rantanen, Harris and Leveille ^{36 )}.

Fig. 2.

(Colour online) Adjusted hazards ratio for major CHD events, cardiovascular mortality and total cause mortality according to sarcopenic obesity groups (defined according to waist circumference (>102 cms) and mid-arm muscle circumference (lowest 2 quintiles)). Hazards ratios adjusted for age, smoking status, alcohol intake, physical activity and social class. The British Regional Heart Study. Data extracted from Atkins et al.⁽ Reference Atkins, Whincup and Morris ^{43 )}.

Sarcopenic obesity and cardiovascular risk factors

Numerous studies have examined the association of sarcopenic obesity with established cardiovascular risk factors. Several cross-sectional studies in Korean populations of older adults have found that sarcopenic obese individuals had the worse cardiovascular risk profile. Sarcopenic obesity (based on skeletal muscle assessed by DXA and obesity measured by either computerised tomography, DXA, BMI or WC) was associated with lower cardiorespiratory fitness, higher fasting glucose levels, a higher risk of hypertension, dyslipidaemia and insulin resistance, and up to an 8-fold increase in risk of the metabolic syndrome compared with non-sarcopenic, non-obese⁽ Reference Kim, Park and Lim ^{48 –} Reference Kim, Park and Kim ^{55 )}. Similar findings were reported in a community-dwelling sample of Taiwanese older adults; sarcopenic obesity (defined by BIA-measured muscle mass and BMI) was associated with the highest risk of metabolic syndrome⁽ Reference Lu, Yang and Chang ^{56 )}. In a large cross-sectional analysis of over 14 000 adults from the National Health and Nutrition Examination Survey III, the sarcopenic obese group (defined by BIA-measured muscle mass and BMI) had the highest risk of insulin resistance and dysglycaemia⁽ Reference Srikanthan, Hevener and Karlamangla ^{57 )}. However, not all studies have shown sarcopenic obese individuals to have the worst profile and some cross-sectional studies have suggested that obese older adults may have higher cardiovascular risk factors than sarcopenic obese subjects⁽ Reference Baumgartner, Wayne and Waters ^{58 –} Reference dos Santos, Gadelha and Safons ^{60 )}. Conflicting results have also been shown regarding the relationship between inflammatory markers and sarcopenic obesity. While some cross-sectional studies have shown sarcopenic obese adults to have the highest level of inflammatory markers as measured by C-reactive protein⁽ Reference Batsis, Mackenzie and Barre ^{47 ,} Reference Schrager, Metter and Simonsick ^{61 )} others have found no found no significant interactions between sarcopenia and obesity with C-reactive protein⁽ Reference Cesari, Kritchevsky and Baumgartner ^{62 )}.

Sarcopenic obesity and CVD

Despite evidence on the relationships between sarcopenic obesity and cardiovascular risk factors, the association between sarcopenic obesity and CVD has been less studied. Two cross-sectional studies reported that older adults with sarcopenic obesity (based on appendicular skeletal muscle mass and per cent body fat from DXA) did not show significantly higher prevalence of CVD compared with non-sarcopenic, non-obese adults⁽ Reference Baumgartner, Wayne and Waters ^{58 ,} Reference Chin, Rhee and Chon ^{63 )}. Prospective studies examining the association between sarcopenic obesity and CVD are limited and we have identified only two such studies to date. In the Cardiovascular Health Study, a large prospective study of community-dwelling older men and women (age ⩾ 65 years), sarcopenic obesity based on WC and muscle strength, was associated with the highest risk of CVD and congestive heart failure. Sarcopenic obese adults showed a 23 % increase in risk of CVD and a 42 % increase in risk of congestive heart failure over 8 years of follow-up compared with non-obese non-sarcopenic subjects⁽ Reference Stephen and Janssen ^{44 )}. However, the risk of CVD events was not significantly greater in the sarcopenic obese group, when defined using BIA-measured muscle mass, implying that muscle strength may be more important than muscle mass. These results are broadly comparable with a prospective study of older men (age 60–79 years), from the British Regional Heart Study which also showed no excess risk of CHD events (fatal or non-fatal myocardial infarction) in sarcopenic obese men (defined by WC and mid-arm muscle circumference) over 11 years of follow-up (Fig. 2)⁽ Reference Atkins, Whincup and Morris ^{43 )}. Sarcopenic obese men showed increased risk of CVD mortality but risk of CVD mortality was similar to those with sarcopenia or obesity alone. However, this study did not consider muscle strength in defining sarcopenic obesity. Overall, findings from cross-sectional and prospective studies do not provide strong evidence for a synergistic effect of sarcopenic obesity on risk of developing CVD.

The inconsistencies in the observed associations of sarcopenic obesity with cardiovascular risk and mortality risk may be explained by the large variability between studies in their sample size and population characteristics, and also due to the heterogeneity in the definition and classification of sarcopenic obesity. To further assess the association between sarcopenic obesity and risk of CVD and mortality there is a need for larger, prospective studies to be carried out which use a standardised definition of sarcopenic obesity.

The obesity paradox in those with CHD

Despite the potential adverse effects of overweight and obesity on CVD risk factors and incidence, numerous investigators and meta-analysis of studies in cohorts with CHD have shown increased survival in those who are overweight or obese⁽ Reference Oreopoulos, Padwal and Kalantar-Zadeh ^{6 –} Reference Lavie, de Schutter and Patel ^{10 )} when defined by high BMI, the obesity paradox, although this has not been seen in all studies after adjustment for confounders⁽ Reference Zeller, Steg and Ravisy ^{64 ,} Reference Kanhai, Kapelle and van der Graaf ^{65 )}. In the RICO Survey, a study of over 2000 patients with acute myocardial infarction, an inverse association was seen between BMI and mortality but this was attenuated after adjustment for factors associated with survival, including prior myocardial infarction, hypertension, diabetes, hyperlipidaemia, smoking and left ventricular ejection fraction⁽ Reference Zeller, Steg and Ravisy ^{64 )}. In the Secondary Manifestations of ARTerial disease study no association was seen at all between BMI and mortality in patients with CVD ⁽ Reference Kanhai, Kapelle and van der Graaf ^{65 )}. Part of the explanation for the obesity paradox may be due to the use of BMI to define obesity, which does not take into account lean muscle mass. BMI is a poor marker of body fat and does not distinguish between fat and lean body mass which has been associated with increased mortality⁽ Reference Allison, Zhu and Plankey ^{13 )}. Abdominal obesity has been suggested as a better marker of obesity risk. Indeed, meta-analysis of cohorts in CHD has shown positive associations between WC and mortality⁽ Reference Coutinho, Goel and Corrêa de Sá ^{66 )}. However, few population studies to date have examined the possible role of lean muscle mass in explaining the obesity paradox in those with CHD. In a recent report from the British Regional Heart Study, it was shown that the vast majority of men with normal body weight had low muscle mass and the prevalence of low muscle mass increased appreciably in the presence of CHD and heart failure (HF) (Table 1). Low muscle mass was shown to be associated with increased mortality in those without HF irrespective of CHD status. The lower risk associated with overweight and obesity in those with CHD was attenuated after adjustment for muscle mass suggesting that the lower risk of mortality associated with excess body weight in men with CHD without HF appears largely associated with higher muscle mass (Table 1)⁽ Reference Wannamethee, Shaper and Whincup ^{67 )}. The inverse association seen in those with HF persisted after adjustment for muscle mass. Although one study showed the obesity paradox to persist in those with CHD even after taking into account muscle mass⁽ Reference Angerås, Albertsson and Karason ^{8 )}, patients with CHD in that study included those with HF. Thus the persistence of lower risk associated with obesity may have been due to the high prevalence of patients with HF in this group. In HF, the obesity paradox may be driven by the deleterious effects of cardiac cachexia (wasting) reflecting the combined loss of muscle and adipose tissue⁽ Reference Marques and Langouche ^{68 )}. Cardiac cachexia observed in end stage HF is associated with a decrease of fat mass in addition to a reduced lean muscle mass⁽ Reference Anker and Rauchhaus ^{69 )}. It has also been postulated that several physiologic mechanisms may explain the protective effect of a higher BMI on mortality⁽ Reference Dorner and Rieder ^{11 )}. NT-proBNP levels, a marker of cardiac dysfunction, are lower in overweight and obese patients; lower NT-proBNP predicts lower mortality⁽ Reference Coutinho, Goel and Corrêa de Sá ^{66 ,} Reference Clerico, Giannoni and Vittorini ^{70 )}. Another possible explanation for the obesity paradox directly involves the functions of adipose tissue⁽ Reference Dorner and Rieder ^{11 ,} Reference Coutinho, Goel and Corrêa de Sá ^{66 )}. Adipose tissue produces leptin which experimental studies suggest may have protective effects in HF⁽ Reference McGaffin, Moravec and Mctiernan ^{71 )} and adiponectin is decreased in obesity; lower concentrations of adiponectin have been associated with lower mortality in patients with CHD or HF⁽ Reference Kistorp, Faber and Galatius ^{72 –} Reference Wannamethee, Whincup and Lennon ^{74 )}.

Table 1.

BMI and adjusted hazards ratio (95% CI) total mortality in men by CHD and heart failure (HF) status

Model 1=adjusted for age, smoking, alcohol intake, social class, physical activity, prevalent diabetes and stroke.

Model 2=Model 1 + low muscle mass.

Low muscle mass defined as lowest quartile of mid-arm muscle circumference.

Source: adapted from Wannamethee et al. ⁽ Reference Wannamethee, Shaper and Whincup ^{67 )}.

The association between WC and mortality in CHD patients in contrast to BMI have found to be positive⁽ Reference Wannamethee, Shaper and Whincup ^{67 )} or null⁽ Reference Kanhai, Kapelle and van der Graaf ^{65 )}. The positive association between central adiposity and mortality in those with CHD is in keeping with the findings that the inverse association seen for BMI in those with CHD may reflect reduced muscle mass.