The present study reports evidence for an association between body composition markers assessed by DXA and the concentration of a number of pro- and anti-inflammatory parameters as well as adipose-related hormones. Although DXA does not provide a direct measurement of FM, LM, and BMC, it is widely used for BC assessment in both clinical and research settings, because of its good values of accuracy and precision, large availability, low costs, low radiation dose, and good correlations with BC measurements obtained by CT and MRI [21, 23,24,25].
Although several studies have shown specific associations with central and/or peripheral fat mass, BMI and waist circumference, and inflammatory molecules such as IL-1Rα, IL-6, and IL-6-sR [26,27,28], we did not find any correlation with the body composition markers and indexes studied and circulating levels of a series of pro- and anti-inflammatory molecules such as IL-6, pentraxin 3, IL-10, TGF-β1, TNF-α, IL-6Rα, gp130, TNF-α-R1, and TNF-α-R2. This difference could be explained by the size of the cohort used, by the technique used to identify body composition and many other factors.
As expected, major differences exist between BC characteristics in elderly women and men. Elderly women have higher fat mass than men while men have higher lean mass and bone content than elderly women. Sex dimorphism in total body composition is present at birth and continues through adulthood . Men maintain their lean mass into the fifth decade of life but then begin to lose muscle mass due to both hormonal changes, decline in activity levels, low protein diet, reduced blood flow, and decreased nerve conduction. Women show a similar decline in lean mass, but they often show greater gains in fatness , even when weight is stable . Such changes continue into old age [4, 31].
Among the adipose-related markers, ghrelin, which is considered an anti-inflammatory molecule, is negatively associated with fat mass in women but not in men, while it is positively associated with SMI in both sexes. Ghrelin levels are reported to fall in obesity, with concentration influenced principally by changes in energy balance. Insulin, in particular, may play an important role in the decrease of ghrelin levels after meals . Even if BMI and insulin are not different between sexes, women have greater fat mass than men, and this could explain the different associations found. Also when comparing the ghrelin levels among the five clusters previously identified by the authors  which differ for BMI and fat mass and lean mass and bone density, a sex difference emerged. In women, the levels of ghrelin decrease as BMI clusters increase from normal weight to low obesity clusters.
In both men and women, leptin is positively associated with fat mass, lean mass, and bone mass markers, while it is negatively associated with SMI. Leptin is a classic adipokine that is secreted by adipocytes, and it increases with weight gain and decreases with weight loss and is also considered as a pro-inflammatory marker . Recent studies have reported, however, that leptin is also produced by skeletal muscle [34, 35] as well as bone cells . Leptin treatment increases muscle mass and decreases the expression of atrophy-related factors such as myostatin, muscle RING-finger protein-1 (MuRF1), and muscle atrophy F-box (MAFbx) in muscle  without any change with age. More recent studies show that the effects of leptin on the skeleton are quite complex, and that leptin deficiency is associated with low bone mass primarily due to reduced cortical bone [38, 39]. Central infusions of leptin in leptin-deficient ob/ob mice actually increase cortical bone formation and total bone mass . Individuals with osteoporosis have reduced levels of leptin in the bone marrow microenvironment .
Adiponectin, together with leptin, is able to regulate energy homeostasis. Low levels of adiponectin, that is considered an anti-inflammatory mediator, have been found in obesity and insulin resistance . In our cohort, both men and women show an inverse relationship with fat and lean mass markers. Our results are in agreement with a recent paper by Baker and colleagues  showing that in elderly, high levels of serum adiponectin are correlated with low BMI, fat, and lean mass BC markers. Moreover, adiponectin levels decrease as clusters increase BMI in both sexes. However, it is interesting to note that in women when comparing clusters with similar BMI (25.09 and 26.62, respectively), adiponectin levels are higher in the overweight group with higher levels of fat and bone mass and lower levels of lean mass. In men, the levels of adiponectin are generally not different among the six clusters.
It has been reported that an increase in fat mass is correlated with markers of inflammation among community-dwelling individuals older than 65 years [26, 27]. The mechanisms inducing obesity-related inflammation are not completely understood; however, the expansion of adipose tissue in response to a positive energy balance may play a major role. When adipose tissue expands, it leads to the activation of macrophages which secrete inflammatory cytokines including TNF-α and IL-6 . In addition, leptin together with resistin could also function as a pro-inflammatory molecule in the presence of obesity , while adiponectin and ghrelin have anti-inflammatory properties. In particular, adiponectin is known to inhibit inflammation by blocking NF-kB activation and reducing such cytokines as TNF-α, IL-6, and IL-18 [45, 46]. Moreover, adiponectin may also play a pro-inflammatory role in arthritic joints by promoting COX2 expression and the synthesis of PGE2, which increases inflammation and pain . Through the elaboration of anti- and pro-inflammatory adipokines that enter the systemic circulation, adipose tissue plays a critical role in regulating the inflammatory response in the setting of calorie restriction, obesity, and aging. However, it is possible that the association with inflammatory markers differs by sex and by adipose tissue location. The most commonly measured inflammatory proteins in nutrition surveys are CRP, which is a measure of acute inflammation, and AGP, which is a measure of chronic inflammation . In our study, CRP correlated with fat mass in both sexes, but only in women with LMI and ALMI. CRP and AGP levels gradually increased as the BMI increased in the clusters in women. It has been shown that the effects of aging on the human immune system are significantly different in men and women, showing a stronger pro-inflammatory response in women . Even if in our study any difference emerged in CRP concentrations between men and women (median 0.84 mg/l and 0.87 mg/l, respectively), women have a significantly higher concentration of AGP compared to men (median 0.67 g/l and 0.61 g//l, respectively). Indeed, AGP was positively correlated with fat markers and LMI only in women. Hemoglobin levels were also significantly lower in women than in men (median 13.7 g/dl and 14.9 g/dl, respectively), and these lower levels of iron could possibly further contribute to the different inflammatory status  between men and women. CRP acts as a positive and albumin as a negative acute-phase reactant . This seems to provide a link to the already mentioned, slightly increased inflammatory state in elderly women. In this context, it is interesting that several clinical studies could demonstrate a link between the specific pattern of increased CRP and decreased albumin concentrations with sarcopenia, frailty, and vascular and non-vascular mortality in elderly subjects [50,51,52].
Among the body composition markers, SMI associates differently from the others with the adipose-related and inflammatory markers analyzed in this study. SMI represents a marker of sarcopenia, together with ALMI [53, 54]. In our cohort, the associations of SMI with adipose-related and inflammatory markers studied are always discordant in both women and men, the only exception being the positive correlation with albumin levels in men. In particular, in both women and men, SMI correlated positively with ghrelin, which is considered an anti-inflammatory molecule, but negatively with leptin, CRP, and AGP, which are considered pro-inflammatory markers. As inflammation is thought to have a role in age-related sarcopenia , the results obtained with SMI are more consistent with respect to those obtained with ALMI when both are considered as markers of sarcopenia. These results fit with the open debate on the use of optimal quantitative markers of sarcopenia and the role of imaging [53, 56]. Moreover, SMI is inversely correlated with BMI and fat mass markers and positively with ALMI but not with LM and LMI, while ALMI is positively correlated with BMI and fat mass markers and also with LM and LMI. These results showed that it is likely that ALMI still represents the general lean mass instead of being a marker of sarcopenia; however, further studies are needed to verify this hypothesis. These last results could be of help in the prevention of sarcopenia.