Obesity and body composition in RA

Obesity can increase the risk of RA development . Obesity induces a proinflammatory state that is mediated by the overproduction of cytokines and adipose tissue-derived proinflammatory molecules known as adipokines . Obese patients also experience increased levels of estrogen and androgen as well as reduced vitamin D concentrations. Cytokines, the adipokine leptin, sex hormones, and reduced vitamin D concentrations are implicated in the pathophysiology of RA . As reviewed by Daien and Sellam , recent studies support a role of excess adiposity in RA development. In an US retrospective case–control study, obesity predicted RA onset (odds ratio (OR) 1.24, 95% CI 1.01 to 1.53) after adjustment for smoking status . In the European Prospective Investigation of Cancer–Norfolk and the Norfolk Arthritis Register Study, obesity associated with incident seronegative RA (hazard ratio (HR) 1.34, 95% CI 1.03 to 1.74) . However, by using data from the prospective Nurses’ Health Study, the risk of developing RA for persons with obesity before the age of 55 years was increased in both seropositive and seronegative patients (HR 1.65, 95% CI 1.34 to 2.05) . In a recent meta-analysis of 11 studies , compared to normal weight subjects, the pooled RRs for RA development were 1.31 (95% CI 1.12 to 1.53) and 1.15 (95% CI 1.03 to 1.29) for obesity and overweight, respectively. The estimated summary RR of incident RA for a five-unit increase in body mass index (BMI) was 1.03 (95% CI 1.01 to 1.05). As obesity rates are increasing worldwide, obesity might increasingly contribute to RA onset.

Whether the prevalence of obesity is increased in patients with RA after diagnosis has similarly been questioned . In a recent investigation, the BMI was significantly larger in 350 RA patients (27.3 kg/m ² ) compared to that in 115,787 non-RA persons from the 2005 Canadian Community Health Survey, cycle 3.1 (26.1 kg/m ² ) . Furthermore, the prevalence of obesity in RA is dependent on population origin. Thus, we found that among African black persons from a developing population, both the mean ± standard deviation (SD) BMI and waist-to-height ratio were markedly lower in RA than in non-RA subjects (29.2 ± 6.6 versus 33.7 ± 8.0, p < 0.0001 and 0.58 ± 0.09 versus 0.62 ± 0.1, p = 0.0003, respectively) .

Most importantly in the present context, RA causes major alterations in body composition through metabolic perturbations . This topic was extensively reviewed in the past . Briefly, patients with RA experience loss of lean mass that has been mostly attributed to inflammation induced increased resting energy expenditure and protein catabolism, and occurs with little or no change in weight as the muscle is replaced by fat. This condition thereby comprises the cooccurrence of reduced lean mass and increased fat mass, and is referred to as “rheumatoid cachexia,” a term initially coined by Roubenhoff and colleagues in 1990 . In relation to this, Stavropoulos et al. demonstrated that for a given body fat content, RA patients have a BMI that is decreased by approximately 2 kg/m ² . In patients with RA, it was proposed that BMI cutoff values for overweight and obesity should be reduced to 23 and 28 kg/m ² , respectively. Rheumatoid cachexia affects up to two-thirds of RA patients. Loss of lean mass increases overall mortality in the general population .

Implicated factors in rheumatoid cachexia include the overproduction of cytokines, particularly TNF-α, as well as decreased bioavailability of insulin-like growth factor. In accordance with this notion, rheumatoid cachexia is associated with anticitrullinated peptide antibody (ACPA) positive and erosive RA, features that represent severe disease ; physical disability in RA is inversely associated with appendicular lean mass and it is directly related to total and appendicular fat mass . Hugo et al. recently reported that nutritional complications of RA including rheumatoid cachexia and MetS were associated with decreased physical activity, the use of glucocorticoids and insulin resistance. Insulin is an anabolic hormone.

As body composition is generally not assessed in RA patients, rheumatoid cachexia typically remains undiagnosed and untreated. Rheumatoid cachexia is associated with severe RA and cardiovascular risk factor profiles including larger total and low-density lipoprotein (LDL) cholesterol concentrations and more prevalent hypertension and MetS . This cardiovascular risk burden is attributable to the increased abdominal adiposity component rather than that of decreased fat mass in rheumatoid cachexia . The potential contribution of rheumatoid cachexia to increased cardiovascular event rates requires further study. Rheumatoid cachexia develops early on in RA, and methotrexate and TNF-α inhibitors may halt its progression but do not reverse it . Metsios and colleagues comprehensively investigated the effects of TNF-α blockade on various components of rheumatoid cachexia, during the 12 weeks after its initiation. An effect on resting energy expenditure was documented, and this was explained by increased physical activity and increased protein intake rather than by antiinflammatory effects.

In contrast to rheumatoid cachexia, frank wasting or classic cachexia with loss of both lean and fat mass as typically found in cancer patients is rarer in RA. This would be expected to result in a reduced BMI, which is not only associated with severe disease but also with increased overall and cardiovascular mortality risk in RA . Interestingly, in the respective studies, excess adiposity was paradoxically associated with reduced overall and unaltered cardiovascular mortality in RA .

Apparently, counterintuitive to the above, excess adiposity influences adversely on other metabolic risk factors in RA. Stavropoulos-Kalinoglu et al. found that BMI associates with an increased prevalence of hypertension, insulin resistance, and MetS in RA. Similarly, patients with RA have a propensity toward abdominal obesity . Among both African black and white women with RA, we showed that BMI was related to systolic and diastolic blood pressure but not related to serum lipid concentrations, whereas abdominal obesity indices were associated with serum lipid concentrations but not with blood pressure values; obesity measures that were associated with plasma glucose concentrations comprised BMI, waist circumference, and waist-to-height ratio (p < 0.05 in multiple confounder-adjusted analysis) . In white patients with RA, BMI was associated with carotid intima-media thickness, while waist-to-hip ratio was related to carotid plaque, and these relationships were explained by other altered metabolic risk factors. By contrast, excess adiposity was unrelated to atherosclerosis among African black women with RA. The recent meta-analysis of 10 studies by Baghdadi et al. that originated in developed populations also showed that obesity was associated with an increased RR of combined CVD morbidity (1.16, 95% CI 1.03 to 1.29) in RA.

Taken together, it is noteworthy that there is seemingly a contrast in reported results across studies on the associations between adiposity status and cardiovascular risk in RA. Thus, although low rather than high BMI predicted cardiovascular mortality , other studies clearly document a direct association of excess adiposity with metabolic risk and atherosclerosis . Wolfe and Michaud recently elucidated this issue in a 12.3-year longitudinal study in 24,535 patients by considering that the recorded mean age in RA cohorts is typically well for >50 years. They confirmed that in all patients, obesity was directly associated with increased metabolic risk factors as well as with MI rates but inversely associated with cardiovascular mortality. However, on stratification by age, the obesity–cardiovascular mortality relationships were reproduced in older patients but reversed to a positive association in those <50 years of age. This protective effect of obesity against cardiovascular mortality in older subjects was also previously demonstrated in the general population . In line with these findings, we reported an obesity-driven direct association of leptin levels with increased endothelial activation in RA patients aged <50 years but not in older patients .

Obesity in RA is associated with increased disease activity and disability but reduced radiographic damage . In this context, adiponectin is reduced in obesity but is implicated in joint damage in patients with RA . Giles et al. showed that the association of excess adiposity with reduced joint damage is explained by low adiponectin concentrations in patients with RA. In addition, although adiponectin concentrations are decreased in RA patients with abdominal obesity and without clinically evident joint damage, they are nevertheless associated with reduced carotid plaque prevalence, an effect that is lost in patients without abdominal obesity and clinically evident joint damage despite the presence of higher adiponectin levels . Taken together, reported data indicate that although adiponectin can enhance joint damage in RA, this adipokine may nevertheless reduce cardiovascular risk in this disease as it does in the general population. Finally, obesity decreases treatment responses to conventional disease-modifying antirheumatic drugs (DMARDs) and anti-TNF-α inhibitors .

Insulin resistance and impaired glucose metabolism in RA

Insulin resistance mediates the development of type 2 diabetes . The prevalence of insulin resistance is increased in RA . In a meta-analysis of 15 studies on traditional cardiovascular risk factors by Boyer et al. , the prevalence of diabetes was increased (OR 1.74, 95% CI 1.22 to 2.50) in RA. Furthermore, an increased frequency of undiagnosed diabetes was reported in RA . Another large electronic medical record database study in the UK reported a 12% increased incidence of diabetes in RA, which was explained by obesity and smoking .

Patients with RA experience clustering of metabolic risk factors . Indeed, excess abdominal adiposity, insulin resistance, low high-density lipoprotein (HDL) cholesterol, high triglyceride concentrations, and hypertension are interrelated in RA. C-reactive protein (CRP) concentrations and disease activity are related to insulin resistance, and disease activity is further associated with decreased β-cell function in RA . Finally, the use of long-term treatment with glucocorticoids, subclinical hypothyroidism, and vitamin D deficiency are independently related to insulin resistance in RA . Vitamin D deficiency may contribute to the enhanced CVD risk associated with RA .

Data from longitudinal intervention studies support the role of systemic inflammation in impaired glucose metabolism among RA patients. Traditional DMARDs reduce insulin resistance in RA . A recent meta-analysis of 8 studies reported a reduction in insulin resistance upon anti-TNF-α inhibition in RA . In one study, anti-TNF-α therapy increased insulin sensitivity in insulin-resistant normal-weight patients but not in obese patients with RA . Another investigation documented that anti-TNF-α therapy not only increased insulin sensitivity but also improved β-cell function and reversed defects in the insulin signaling cascade in insulin-resistant patients with RA .

There has been a changing pattern over time in reported findings on the impact of insulin resistance on CVD in patients with RA. In 2006, we showed an independent relationship between insulin resistance and carotid atherosclerosis as evidenced by plaque presence in RA . In keeping with these findings, in 2007, Chung et al. documented that insulin resistance was independently related to coronary atherosclerosis as estimated by electron-beam computed tomography in RA. In fact, insulin resistance was more consistently associated with atherosclerosis than with applied MetS definitions in RA . By contrast, in 2015, Giles et al. found that insulin resistance was no longer associated with systemic inflammation, neither with the presence or development of increasing coronary artery calcium scores or ultrasound-determined carotid artery intima-media thickness or plaque. The systemic inflammatory burden was much lower in the Gile study than in the 2 previously reported investigations . Thus, this disparity in insulin-CVD risk among studies may represent our increasing success in achieving better RA control leading to a lower systemic inflammatory burden.

Lipids

Altered lipid profiles comprise a pivotal modifiable cardiovascular risk factor in the population at large. In the meta-analysis by Boyer et al. , the prevalence of hypercholesterolemia was not increased in RA. By contrast, HDL cholesterol concentrations were reduced in RA patients compared to controls with a weighted mean difference of −17.72 (95% CI −18.35 to −17.08) mg/dl .

Systemic inflammation has consistent and marked effects on circulating lipid concentrations in RA . Generally, lipid fractions including total, LDL and HDL cholesterol, and apolipoprotein A1 concentrations are reduced by systemic inflammation, as mostly estimated by CRP levels in RA. Yet, patients with active RA are reportedly at increased CVD risk. This is often referred to as the “lipid paradox” in RA . The same process occurs in patients with other diseases associated with inflammation and cachexia such as those with chronic kidney disease. In active RA, the reduction in lipid concentrations can be attributed to increased catabolism by the reticuloendothelial system . In addition to reducing lipid fractions, systemic inflammation alters HDL structure and function . HDL plays a pivotal role in reverse cholesterol transport as it facilitates cholesterol efflux from vascular lesions. In normal health, paraoxonases bound to HDL further reduce atherogenesis by protecting against oxidation of lipids and promoting the beneficial effects of HDL on cholesterol efflux. Because of molecular changes in active RA, HDL loses its antiinflammatory, antioxidant, and atheroprotective properties and becomes proinflammatory .

In support of systemic inflammation-induced low lipid concentrations, several studies have shown that effective treatment of active RA with traditional DMARDs is associated with an increase in total, HDL, and LDL cholesterol levels soon after initiation . In a meta-analysis of 13 prospective studies , the use of anti-TNF-α therapy for 52 weeks in RA was associated with increased concentrations of HDL and total cholesterol but unaltered LDL concentrations; after long-term treatment, triglyceride levels increased and apolipoprotein B/A1 ratios decreased. Interestingly, in a 2-year follow-up study , the use of triple therapy comprising methotrexate, hydroxychloroquine, and sulfasalazine, was associated with higher HDL and lower LDL cholesterol concentrations, and lower total cholesterol–HDL cholesterol ratios than those observed in methotrexate monotherapy and methotrexate plus etanercept combination therapy. A hydroxychloroquine-mediated reduction in hepatic cholesterol synthesis may have explained these disparities . Increased lipid concentrations are mostly reported with the use of the interleukin-6 inhibitor, tocilizumab and the oral Janus kinase inhibitor, tofacitinib. In fact, in a recent meta-analysis of 25 randomized placebo-controlled studies, tocilizumab and tofacitinib increased LDL and HDL concentrations, and tocilizumab further increased total cholesterol levels, whereas changes in lipid concentrations with anti-TNF-α therapy were not significant . Notably, methotrexate and adalimumab also favorably impact on not only lipoprotein function but also macrophage cholesterol handling, and could thereby oppose foam cell formation .

The impact of systemic inflammation on lipid profiles has important implications in CVD risk assessment and its management in patients with RA. Although lipid fractions often change markedly in association with systemic inflammation as well as with effective disease activity reduction with traditional and biological DMARDs and tofacitinib, the total cholesterol–HDL cholesterol ratio remains mostly constant. Hence, current recommendations on CVD prevention include focusing on the latter lipid variable . Nevertheless, lipid concentrations should preferably be assessed when RA disease activity is stable and, reevaluated subsequent to changes in DMARD therapy. Whether the consideration of changes in HDL function can improve CVD risk stratification in RA is currently unknown. TNF-α inhibitors, tocilizumab, and tofacitinib can remodel HDL cholesterol particles from a proinflammatory to an antiinflammatory phenotype .

In sharp contrast to the results obtained in the investigations as discussed above, the recent meta-analysis by Baghdadi et al. revealed a RR of 1.73 (95% CI 1.03 to 2.44) of cardiovascular morbidity associated with hypercholesterolemia in RA patients. Again, this may represent a consequence of the increasingly effective disease activity control that results in a reduced inflammatory burden in more recently treated patients with RA.

Hypertension in RA

Along with altered lipid profiles, hypertension comprises another essential modifiable traditional cardiovascular risk factor in the general population. In the meta-analysis by Boyer et al., published in 2011 , the prevalence of hypertension was unaltered in RA with an OR of 1.09 (95% CI 0.91 to 1.31). However, in a recent investigation of 502,649 UK Biobank participants, the prevalence of hypertension in RA subjects was increased with a covariate-adjusted OR of 1.19 (95% CI 1.21 to 1.27) .

Hypertension is strongly associated with atherosclerosis in RA . In the meta-analysis by Baghdadi et al. , the risk of MI (RR 1.84, 95% CI 1.38 to 2.46) and cardiovascular morbidity (RR 2.24, 95% CI 1.42 to 3.06) associated with hypertension was increased in patients with RA.

Seminal work by Panoulas et al. confirmed a high prevalence of hypertension at 70.5% in RA. More importantly, 39.4% of hypertensive RA patients were undiagnosed and among those treated, 78.2% were inadequately controlled. BMI and prednisone use was independently associated with the presence of hypertension, and BMI and the prevalent CVD were related to uncontrolled hypertension. Other factors implicated in hypertension among RA patients comprise systemic inflammation, physical inactivity, and other antirheumatic agents including nonsteroidal antiinflammatory drugs and leflunomide .

Overall metabolic risk: the MetS in RA

Among six available clinical definitions, the National Cholesterol Education Program–Adult Treatment Program III (NCEP–ATP III) MetS is most frequently applied in RA and non-RA studies . Accordingly, patients are diagnosed with MetS when three or more of the following are present: (1) waist circumference ≥102 cm in white men and ≥90 cm in South Asian men, and ≥88 cm in white women and ≥80 cm in South Asian women; (2) triglycerides ≥150 mg/dl (1.7 mmol/l); (3) HDL cholesterol <40 mg/dl (1.0 mmol/l) in men and <50 mg/dl (1.3 mmol/l) in women; (4) systolic blood pressure ≥130 mm Hg and/or diastolic blood pressure ≥85 mm Hg or drug treatment for hypertension; and (5) fasting plasma glucose ≥100 mg/dl (5.6 mmol/l). In a meta-analysis of 4 cross-sectional plus 8 case–control studies , RA was associated with an increased prevalence of NCEP–ATP III MetS (OR 1.24, 95% CI 1.03 to 1.50). Notably, in subgroup analysis, RA was associated with MetS in American studies (n = 4; OR 1.57, 95% CI 1.25 to 1.97) but not in those performed in Europe (n = 2; OR 1.17, 95% CI 0.84 to 1.64) or Asia (n = 5; OR 1.17, 95% CI 0.84 to 1.64) . These disparities may represent regional differences in lifestyle factors as well as ethnic differences in body composition and associated cardiovascular metabolic risk factors.

High-grade inflammation induces metabolic risk factors including insulin resistance, reduced HDL cholesterol levels, and increased blood pressure in RA . Indeed, active RA and MetS share several characteristics. Furthermore, obesity mediates systemic inflammation . Therefore, a bidirectional reciprocally reinforcing cross-talk between systemic inflammation and metabolic risk would be expected. Evidence confirming this hypothesis was provided by Parra-Salcedo et al. in a 24-month longitudinal study on MetS in patients with early RA treated to target with conventional DMARDs. Cumulative disease activity and high BMI increased the incidence of MetS, whereas incident MetS and high baseline disease activity prevented RA remission. These findings have an important clinical implication: treatment of both disease activity and metabolic risk factors should be considered in improving outcomes in RA.

In cross-sectional studies, the use of glucocorticoids was associated with an increased prevalence of MetS in RA , whereas use of methotrexate and hydroxychloroquine related to less frequent MetS. This suggests that antirheumatic drugs may have different effects on cardiometabolic risk that are partially independent of their antiinflammatory effects.

The potential involvement of adipokines in the pathophysiology and cardiovascular risk associated with inflammatory arthritis has been the topic of many recent reviews. In addition, these molecules do not have applications in clinical medicine yet. Hence, adipokines in relation to inflammatory arthritis are not systematically reviewed in this manuscript.

Obesity and body composition in RA

Obesity can increase the risk of RA development . Obesity induces a proinflammatory state that is mediated by the overproduction of cytokines and adipose tissue-derived proinflammatory molecules known as adipokines . Obese patients also experience increased levels of estrogen and androgen as well as reduced vitamin D concentrations. Cytokines, the adipokine leptin, sex hormones, and reduced vitamin D concentrations are implicated in the pathophysiology of RA . As reviewed by Daien and Sellam , recent studies support a role of excess adiposity in RA development. In an US retrospective case–control study, obesity predicted RA onset (odds ratio (OR) 1.24, 95% CI 1.01 to 1.53) after adjustment for smoking status . In the European Prospective Investigation of Cancer–Norfolk and the Norfolk Arthritis Register Study, obesity associated with incident seronegative RA (hazard ratio (HR) 1.34, 95% CI 1.03 to 1.74) . However, by using data from the prospective Nurses’ Health Study, the risk of developing RA for persons with obesity before the age of 55 years was increased in both seropositive and seronegative patients (HR 1.65, 95% CI 1.34 to 2.05) . In a recent meta-analysis of 11 studies , compared to normal weight subjects, the pooled RRs for RA development were 1.31 (95% CI 1.12 to 1.53) and 1.15 (95% CI 1.03 to 1.29) for obesity and overweight, respectively. The estimated summary RR of incident RA for a five-unit increase in body mass index (BMI) was 1.03 (95% CI 1.01 to 1.05). As obesity rates are increasing worldwide, obesity might increasingly contribute to RA onset.

Whether the prevalence of obesity is increased in patients with RA after diagnosis has similarly been questioned . In a recent investigation, the BMI was significantly larger in 350 RA patients (27.3 kg/m ² ) compared to that in 115,787 non-RA persons from the 2005 Canadian Community Health Survey, cycle 3.1 (26.1 kg/m ² ) . Furthermore, the prevalence of obesity in RA is dependent on population origin. Thus, we found that among African black persons from a developing population, both the mean ± standard deviation (SD) BMI and waist-to-height ratio were markedly lower in RA than in non-RA subjects (29.2 ± 6.6 versus 33.7 ± 8.0, p < 0.0001 and 0.58 ± 0.09 versus 0.62 ± 0.1, p = 0.0003, respectively) .

Most importantly in the present context, RA causes major alterations in body composition through metabolic perturbations . This topic was extensively reviewed in the past . Briefly, patients with RA experience loss of lean mass that has been mostly attributed to inflammation induced increased resting energy expenditure and protein catabolism, and occurs with little or no change in weight as the muscle is replaced by fat. This condition thereby comprises the cooccurrence of reduced lean mass and increased fat mass, and is referred to as “rheumatoid cachexia,” a term initially coined by Roubenhoff and colleagues in 1990 . In relation to this, Stavropoulos et al. demonstrated that for a given body fat content, RA patients have a BMI that is decreased by approximately 2 kg/m ² . In patients with RA, it was proposed that BMI cutoff values for overweight and obesity should be reduced to 23 and 28 kg/m ² , respectively. Rheumatoid cachexia affects up to two-thirds of RA patients. Loss of lean mass increases overall mortality in the general population .

Implicated factors in rheumatoid cachexia include the overproduction of cytokines, particularly TNF-α, as well as decreased bioavailability of insulin-like growth factor. In accordance with this notion, rheumatoid cachexia is associated with anticitrullinated peptide antibody (ACPA) positive and erosive RA, features that represent severe disease ; physical disability in RA is inversely associated with appendicular lean mass and it is directly related to total and appendicular fat mass . Hugo et al. recently reported that nutritional complications of RA including rheumatoid cachexia and MetS were associated with decreased physical activity, the use of glucocorticoids and insulin resistance. Insulin is an anabolic hormone.

As body composition is generally not assessed in RA patients, rheumatoid cachexia typically remains undiagnosed and untreated. Rheumatoid cachexia is associated with severe RA and cardiovascular risk factor profiles including larger total and low-density lipoprotein (LDL) cholesterol concentrations and more prevalent hypertension and MetS . This cardiovascular risk burden is attributable to the increased abdominal adiposity component rather than that of decreased fat mass in rheumatoid cachexia . The potential contribution of rheumatoid cachexia to increased cardiovascular event rates requires further study. Rheumatoid cachexia develops early on in RA, and methotrexate and TNF-α inhibitors may halt its progression but do not reverse it . Metsios and colleagues comprehensively investigated the effects of TNF-α blockade on various components of rheumatoid cachexia, during the 12 weeks after its initiation. An effect on resting energy expenditure was documented, and this was explained by increased physical activity and increased protein intake rather than by antiinflammatory effects.

In contrast to rheumatoid cachexia, frank wasting or classic cachexia with loss of both lean and fat mass as typically found in cancer patients is rarer in RA. This would be expected to result in a reduced BMI, which is not only associated with severe disease but also with increased overall and cardiovascular mortality risk in RA . Interestingly, in the respective studies, excess adiposity was paradoxically associated with reduced overall and unaltered cardiovascular mortality in RA .

Apparently, counterintuitive to the above, excess adiposity influences adversely on other metabolic risk factors in RA. Stavropoulos-Kalinoglu et al. found that BMI associates with an increased prevalence of hypertension, insulin resistance, and MetS in RA. Similarly, patients with RA have a propensity toward abdominal obesity . Among both African black and white women with RA, we showed that BMI was related to systolic and diastolic blood pressure but not related to serum lipid concentrations, whereas abdominal obesity indices were associated with serum lipid concentrations but not with blood pressure values; obesity measures that were associated with plasma glucose concentrations comprised BMI, waist circumference, and waist-to-height ratio (p < 0.05 in multiple confounder-adjusted analysis) . In white patients with RA, BMI was associated with carotid intima-media thickness, while waist-to-hip ratio was related to carotid plaque, and these relationships were explained by other altered metabolic risk factors. By contrast, excess adiposity was unrelated to atherosclerosis among African black women with RA. The recent meta-analysis of 10 studies by Baghdadi et al. that originated in developed populations also showed that obesity was associated with an increased RR of combined CVD morbidity (1.16, 95% CI 1.03 to 1.29) in RA.

Taken together, it is noteworthy that there is seemingly a contrast in reported results across studies on the associations between adiposity status and cardiovascular risk in RA. Thus, although low rather than high BMI predicted cardiovascular mortality , other studies clearly document a direct association of excess adiposity with metabolic risk and atherosclerosis . Wolfe and Michaud recently elucidated this issue in a 12.3-year longitudinal study in 24,535 patients by considering that the recorded mean age in RA cohorts is typically well for >50 years. They confirmed that in all patients, obesity was directly associated with increased metabolic risk factors as well as with MI rates but inversely associated with cardiovascular mortality. However, on stratification by age, the obesity–cardiovascular mortality relationships were reproduced in older patients but reversed to a positive association in those <50 years of age. This protective effect of obesity against cardiovascular mortality in older subjects was also previously demonstrated in the general population . In line with these findings, we reported an obesity-driven direct association of leptin levels with increased endothelial activation in RA patients aged <50 years but not in older patients .

Obesity in RA is associated with increased disease activity and disability but reduced radiographic damage . In this context, adiponectin is reduced in obesity but is implicated in joint damage in patients with RA . Giles et al. showed that the association of excess adiposity with reduced joint damage is explained by low adiponectin concentrations in patients with RA. In addition, although adiponectin concentrations are decreased in RA patients with abdominal obesity and without clinically evident joint damage, they are nevertheless associated with reduced carotid plaque prevalence, an effect that is lost in patients without abdominal obesity and clinically evident joint damage despite the presence of higher adiponectin levels . Taken together, reported data indicate that although adiponectin can enhance joint damage in RA, this adipokine may nevertheless reduce cardiovascular risk in this disease as it does in the general population. Finally, obesity decreases treatment responses to conventional disease-modifying antirheumatic drugs (DMARDs) and anti-TNF-α inhibitors .

Insulin resistance and impaired glucose metabolism in RA

Insulin resistance mediates the development of type 2 diabetes . The prevalence of insulin resistance is increased in RA . In a meta-analysis of 15 studies on traditional cardiovascular risk factors by Boyer et al. , the prevalence of diabetes was increased (OR 1.74, 95% CI 1.22 to 2.50) in RA. Furthermore, an increased frequency of undiagnosed diabetes was reported in RA . Another large electronic medical record database study in the UK reported a 12% increased incidence of diabetes in RA, which was explained by obesity and smoking .

Patients with RA experience clustering of metabolic risk factors . Indeed, excess abdominal adiposity, insulin resistance, low high-density lipoprotein (HDL) cholesterol, high triglyceride concentrations, and hypertension are interrelated in RA. C-reactive protein (CRP) concentrations and disease activity are related to insulin resistance, and disease activity is further associated with decreased β-cell function in RA . Finally, the use of long-term treatment with glucocorticoids, subclinical hypothyroidism, and vitamin D deficiency are independently related to insulin resistance in RA . Vitamin D deficiency may contribute to the enhanced CVD risk associated with RA .

Data from longitudinal intervention studies support the role of systemic inflammation in impaired glucose metabolism among RA patients. Traditional DMARDs reduce insulin resistance in RA . A recent meta-analysis of 8 studies reported a reduction in insulin resistance upon anti-TNF-α inhibition in RA . In one study, anti-TNF-α therapy increased insulin sensitivity in insulin-resistant normal-weight patients but not in obese patients with RA . Another investigation documented that anti-TNF-α therapy not only increased insulin sensitivity but also improved β-cell function and reversed defects in the insulin signaling cascade in insulin-resistant patients with RA .

There has been a changing pattern over time in reported findings on the impact of insulin resistance on CVD in patients with RA. In 2006, we showed an independent relationship between insulin resistance and carotid atherosclerosis as evidenced by plaque presence in RA . In keeping with these findings, in 2007, Chung et al. documented that insulin resistance was independently related to coronary atherosclerosis as estimated by electron-beam computed tomography in RA. In fact, insulin resistance was more consistently associated with atherosclerosis than with applied MetS definitions in RA . By contrast, in 2015, Giles et al. found that insulin resistance was no longer associated with systemic inflammation, neither with the presence or development of increasing coronary artery calcium scores or ultrasound-determined carotid artery intima-media thickness or plaque. The systemic inflammatory burden was much lower in the Gile study than in the 2 previously reported investigations . Thus, this disparity in insulin-CVD risk among studies may represent our increasing success in achieving better RA control leading to a lower systemic inflammatory burden.

Lipids

Altered lipid profiles comprise a pivotal modifiable cardiovascular risk factor in the population at large. In the meta-analysis by Boyer et al. , the prevalence of hypercholesterolemia was not increased in RA. By contrast, HDL cholesterol concentrations were reduced in RA patients compared to controls with a weighted mean difference of −17.72 (95% CI −18.35 to −17.08) mg/dl .

Systemic inflammation has consistent and marked effects on circulating lipid concentrations in RA . Generally, lipid fractions including total, LDL and HDL cholesterol, and apolipoprotein A1 concentrations are reduced by systemic inflammation, as mostly estimated by CRP levels in RA. Yet, patients with active RA are reportedly at increased CVD risk. This is often referred to as the “lipid paradox” in RA . The same process occurs in patients with other diseases associated with inflammation and cachexia such as those with chronic kidney disease. In active RA, the reduction in lipid concentrations can be attributed to increased catabolism by the reticuloendothelial system . In addition to reducing lipid fractions, systemic inflammation alters HDL structure and function . HDL plays a pivotal role in reverse cholesterol transport as it facilitates cholesterol efflux from vascular lesions. In normal health, paraoxonases bound to HDL further reduce atherogenesis by protecting against oxidation of lipids and promoting the beneficial effects of HDL on cholesterol efflux. Because of molecular changes in active RA, HDL loses its antiinflammatory, antioxidant, and atheroprotective properties and becomes proinflammatory .

In support of systemic inflammation-induced low lipid concentrations, several studies have shown that effective treatment of active RA with traditional DMARDs is associated with an increase in total, HDL, and LDL cholesterol levels soon after initiation . In a meta-analysis of 13 prospective studies , the use of anti-TNF-α therapy for 52 weeks in RA was associated with increased concentrations of HDL and total cholesterol but unaltered LDL concentrations; after long-term treatment, triglyceride levels increased and apolipoprotein B/A1 ratios decreased. Interestingly, in a 2-year follow-up study , the use of triple therapy comprising methotrexate, hydroxychloroquine, and sulfasalazine, was associated with higher HDL and lower LDL cholesterol concentrations, and lower total cholesterol–HDL cholesterol ratios than those observed in methotrexate monotherapy and methotrexate plus etanercept combination therapy. A hydroxychloroquine-mediated reduction in hepatic cholesterol synthesis may have explained these disparities . Increased lipid concentrations are mostly reported with the use of the interleukin-6 inhibitor, tocilizumab and the oral Janus kinase inhibitor, tofacitinib. In fact, in a recent meta-analysis of 25 randomized placebo-controlled studies, tocilizumab and tofacitinib increased LDL and HDL concentrations, and tocilizumab further increased total cholesterol levels, whereas changes in lipid concentrations with anti-TNF-α therapy were not significant . Notably, methotrexate and adalimumab also favorably impact on not only lipoprotein function but also macrophage cholesterol handling, and could thereby oppose foam cell formation .

The impact of systemic inflammation on lipid profiles has important implications in CVD risk assessment and its management in patients with RA. Although lipid fractions often change markedly in association with systemic inflammation as well as with effective disease activity reduction with traditional and biological DMARDs and tofacitinib, the total cholesterol–HDL cholesterol ratio remains mostly constant. Hence, current recommendations on CVD prevention include focusing on the latter lipid variable . Nevertheless, lipid concentrations should preferably be assessed when RA disease activity is stable and, reevaluated subsequent to changes in DMARD therapy. Whether the consideration of changes in HDL function can improve CVD risk stratification in RA is currently unknown. TNF-α inhibitors, tocilizumab, and tofacitinib can remodel HDL cholesterol particles from a proinflammatory to an antiinflammatory phenotype .

In sharp contrast to the results obtained in the investigations as discussed above, the recent meta-analysis by Baghdadi et al. revealed a RR of 1.73 (95% CI 1.03 to 2.44) of cardiovascular morbidity associated with hypercholesterolemia in RA patients. Again, this may represent a consequence of the increasingly effective disease activity control that results in a reduced inflammatory burden in more recently treated patients with RA.

Hypertension in RA

Along with altered lipid profiles, hypertension comprises another essential modifiable traditional cardiovascular risk factor in the general population. In the meta-analysis by Boyer et al., published in 2011 , the prevalence of hypertension was unaltered in RA with an OR of 1.09 (95% CI 0.91 to 1.31). However, in a recent investigation of 502,649 UK Biobank participants, the prevalence of hypertension in RA subjects was increased with a covariate-adjusted OR of 1.19 (95% CI 1.21 to 1.27) .

Hypertension is strongly associated with atherosclerosis in RA . In the meta-analysis by Baghdadi et al. , the risk of MI (RR 1.84, 95% CI 1.38 to 2.46) and cardiovascular morbidity (RR 2.24, 95% CI 1.42 to 3.06) associated with hypertension was increased in patients with RA.

Seminal work by Panoulas et al. confirmed a high prevalence of hypertension at 70.5% in RA. More importantly, 39.4% of hypertensive RA patients were undiagnosed and among those treated, 78.2% were inadequately controlled. BMI and prednisone use was independently associated with the presence of hypertension, and BMI and the prevalent CVD were related to uncontrolled hypertension. Other factors implicated in hypertension among RA patients comprise systemic inflammation, physical inactivity, and other antirheumatic agents including nonsteroidal antiinflammatory drugs and leflunomide .

Overall metabolic risk: the MetS in RA

Among six available clinical definitions, the National Cholesterol Education Program–Adult Treatment Program III (NCEP–ATP III) MetS is most frequently applied in RA and non-RA studies . Accordingly, patients are diagnosed with MetS when three or more of the following are present: (1) waist circumference ≥102 cm in white men and ≥90 cm in South Asian men, and ≥88 cm in white women and ≥80 cm in South Asian women; (2) triglycerides ≥150 mg/dl (1.7 mmol/l); (3) HDL cholesterol <40 mg/dl (1.0 mmol/l) in men and <50 mg/dl (1.3 mmol/l) in women; (4) systolic blood pressure ≥130 mm Hg and/or diastolic blood pressure ≥85 mm Hg or drug treatment for hypertension; and (5) fasting plasma glucose ≥100 mg/dl (5.6 mmol/l). In a meta-analysis of 4 cross-sectional plus 8 case–control studies , RA was associated with an increased prevalence of NCEP–ATP III MetS (OR 1.24, 95% CI 1.03 to 1.50). Notably, in subgroup analysis, RA was associated with MetS in American studies (n = 4; OR 1.57, 95% CI 1.25 to 1.97) but not in those performed in Europe (n = 2; OR 1.17, 95% CI 0.84 to 1.64) or Asia (n = 5; OR 1.17, 95% CI 0.84 to 1.64) . These disparities may represent regional differences in lifestyle factors as well as ethnic differences in body composition and associated cardiovascular metabolic risk factors.

High-grade inflammation induces metabolic risk factors including insulin resistance, reduced HDL cholesterol levels, and increased blood pressure in RA . Indeed, active RA and MetS share several characteristics. Furthermore, obesity mediates systemic inflammation . Therefore, a bidirectional reciprocally reinforcing cross-talk between systemic inflammation and metabolic risk would be expected. Evidence confirming this hypothesis was provided by Parra-Salcedo et al. in a 24-month longitudinal study on MetS in patients with early RA treated to target with conventional DMARDs. Cumulative disease activity and high BMI increased the incidence of MetS, whereas incident MetS and high baseline disease activity prevented RA remission. These findings have an important clinical implication: treatment of both disease activity and metabolic risk factors should be considered in improving outcomes in RA.

In cross-sectional studies, the use of glucocorticoids was associated with an increased prevalence of MetS in RA , whereas use of methotrexate and hydroxychloroquine related to less frequent MetS. This suggests that antirheumatic drugs may have different effects on cardiometabolic risk that are partially independent of their antiinflammatory effects.

The potential involvement of adipokines in the pathophysiology and cardiovascular risk associated with inflammatory arthritis has been the topic of many recent reviews. In addition, these molecules do not have applications in clinical medicine yet. Hence, adipokines in relation to inflammatory arthritis are not systematically reviewed in this manuscript.