OUP user menu

Adipokines—targeting a root cause of cardiometabolic risk

A. Bakhai
DOI: http://dx.doi.org/10.1093/qjmed/hcn066 767-776 First published online: 11 June 2008


Obesity often co-presents with other cardiometabolic risk factors such as dyslipidaemia, insulin resistance and hypertension. Less well appreciated is that dysregulation of adipokine production by excess adipose tissue also promotes a state of low-level systemic chronic inflammation and a prothrombotic state, implicated in the development of both atherosclerosis and subsequently cardiovascular events. Lifestyle modification and pharmacological therapy can reduce cardiometabolic risk, a benefit that may be partly due to their effects on adipokine levels.


Despite advances in management, cardiovascular disease (CVD) remains one of the leading causes of death, accounting for 37% of deaths in the UK in 2004.1 While considerable strides have been made to increase revascularization rates and the management of dyslipidaemia, hypertension and diabetes, obesity-related diseases are offsetting gains made. Abdominal obesity is a strong independent risk factor for cardiovascular mortality and also increases the incidence of other cardiovascular risk factors and type 2 diabetes.2 Targeting this epidemic is the next important challenge for the medical profession and public at large.

There are several aspects to the contributions of obesity to CVD. Increased release of non-esterified fatty acids (NEFAs) leads to accumulation of fatty acids in cells other than adipocytes, such as muscle and the liver. This ‘lipotoxicity’ disrupts normal insulin signalling leading to insulin resistance.3 In conjunction, abnormal production of protein molecules, called adipokines, by excess and expanded adipose tissue contribute to cardiovascular risk and insulin resistance as well as being mediators of several pathologies, including depression and sleep apnoea. This review will focus on the role of adipokines in cardiometabolic risk and explore management options that reduce the levels of adipokines, either directly or via weight loss.

Adipose tissue, adipokines and inflammation in obesity

Rather than being a passive energy store, adipose tissue is now recognized as being an important secretory organ, producing a range of bioactive proteins called adipokines. These adipokines have several roles including regulating appetite and energy balance, lipid metabolism, blood pressure, insulin sensitivity, haemostasis and angiogenesis (Table 1).4 The term adipokine refers to the molecules produced by adipose tissue, and although many of these molecules are also produced by other cells and tissues; true adipokines, such as leptin and adiponectin, are produced exclusively by adipose tissue.

View this table:
Table 1

Some adipokines and their physiological roles4,7

AdiponectinAnti-inflammatory Insulin sensitizer
LeptinRegulates appetite and energy expenditure
PAI-1Contributes to the prothrombotic state proinflammatory
TNF-αImpairs the insulin signalling cascade, inducing insulin resistance Contributes to the proinflammatory state
IL-6Stimulates CRP release from the liver Causes insulin resistance Contributes to the proinflammatory state
AngiotensinogenContributes to hypertension
ResistinInduces insulin resistance Proinflammatory

Increased abdominal obesity with a predominance of visceral fat, as seen in upper-body obesity, increases the risk of cardiovascular or metabolic disease.5 This appears to be due to differences in metabolism and adipokine secretion between visceral and subcutaneous fat. Release of NEFAs is more rapid from visceral fat than subcutaneous fat due to greater lipolytic activity, especially in obese subjects, which contributes to NEFA levels in the systemic circulation.6 Increased release of NEFAs into the portal circulation stimulates hepatic glucose production and reduces hepatic insulin clearance, ultimately resulting in insulin resistance, hyperinsulinaemia, hyperglycaemia and non-alcoholic fatty liver disease (NAFLD).6,7 Visceral adipose tissue secretes CETP, plasminogen activator inhibitor (PAI-1), angiotensinogen, adiponectin and IL-6 at greater levels than subcutaneous fat, which produces more acylation stimulating protein and leptin.8 Abnormal production of these adipokines by expanded visceral fat during obesity contributes to a proinflammatory state.

Obesity is characterized by a state of chronic low-grade inflammation, indicated by increased plasma levels of several markers of inflammation, including interleukin-6 (IL-6), tumour necrosis factor-α (TNF-α) and C-reactive protein (CRP).9 It is increasingly evident that this state of inflammation may contribute to the health problems associated with obesity, such as dyslipidaemia, insulin resistance and atherosclerosis (Figure 1).10 The link between obesity and inflammation may lie in the observation that levels of the adipokine resistin, produced by macrophages in the stromal vascular region, correlate reasonably with levels of proinflammatory marker CRP.11 The role of resistin continues to be elucidated.

Figure 1.

Altered expression of adipokines associated with obesity induce a state of low-level systemic inflammation and dyslipidaemia leading eventually to atherosclerosis. Adapted with permission from Lyon CJ et al.23 © the Endocrine Society.

Several proinflammatory cytokines are also secreted by adipose tissue, such as TNF-α, IL-6 and leptin, as well as acute-phase proteins, such as PAI-1 and CRP.4 Plasma levels of these adipokines increase along with increasing adipose mass. Although the evidence for low-level expression of CRP in adipose tissue is inconclusive, its expression in the liver is further up-regulated by IL-6, which is secreted in increased amounts in obesity.12

CRP measured appropriately is emerging as a strong independent predictor of cardiovascular events. Several large studies have found a concentration-dependent relationship between CRP levels and risk of coronary heart disease (CHD).13,14 In the Women's Health Study, CRP was the strongest predictor of future cardiovascular events of 12 markers examined, including low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C) and total cholesterol.14 Adipokines not only, therefore, have direct effects on vessels but also mediate some of their effects by raising CRP.

Visfatin is an adipokine identified as being predominantly produced by abdominal adipose tissue.15 Plasma levels of visfatin are closely related to white adipose tissue accumulation and it was originally shown to have insulin-mimetic properties, although this is now under debate. Levels of the adipokine visfatin are higher in overweight or obese patients with additional cardiovascular risk factors, such as increased waist circumference, blood pressure and triglyceride levels, than in patients without these factors.16 Visfatin is considered to be a proinflammatory adipokine and has been shown to upregulate expression of TNF-α and IL-6.17

In contrast to other adipokines, levels of the anti-inflammatory adipokine adiponectin are reduced in obese subjects.18 This may further exacerbate the state of low-grade systemic inflammation associated with obesity. Indeed, low plasma levels of adiponectin have been associated with increased risk of myocardial infarction (MI).19 In addition, levels of adiponectin in plasma and adipose tissue negatively correlate with CRP concentrations.20

Adipose tissue is also a significant source of cholesterol ester transfer protein (CETP), plasma concentrations of which are increased in obese subjects.21 CETP activity is a major determinant of lipoprotein composition, which will be discussed further below.

Adipokines and atherosclerosis

Adipose tissue produces and secretes inflammatory factors that are known to play important roles in the atherosclerotic process. These include TNF-α, leptin, PAI-1, IL-6, resistin and angiotensinogen.22 Expression and plasma levels of these adipokines increase in proportion to adiposity.

Adipokines enhance the attachment and migration of monocytes into the vessel wall and their conversion into macrophages, where they phagocytose oxidized LDL and form lipid-laden foam cells. This is a key early stage in the development of atherosclerosis—as the foam cells accumulate in the vessel wall they form fatty streaks that ultimately develop into atherosclerotic plaques.23 TNF-α induces expression of nuclear factor-κB (NF-κB), which causes the expression of adhesion molecules, macrophage chemoattractive protein-1 and macrophage colony-stimulating factor in the endothelial and vascular smooth muscle cells.23

Angiotensinogen, the precursor to angiotensin II, also induces expression of several vascular adhesion molecules through activation of NF-κB.24 In addition, it enhances metabolism of nitric oxide into free radical species, which damage vascular tissue and remove the protective effect of nitric oxide against macrophage adhesion and accumulation. Leptin acts on macrophages by stimulating the release of macrophage colony-stimulating factor and causing macrophages to accumulate cholesterol, especially in the presence of glucose.25 PAI-1 inhibits the breakdown of fibrin clots, thereby promoting thrombus formation when a plaque ruptures, and is a major contributor to the prothrombotic state seen in patients with insulin resistance.26 Resistin has also been reported to increase the expression of adhesion molecules and endothelin-1, a molecule linked to endothelial dysfunction and CVD.27 In addition, elevated plasma levels of resistin have been found to correlate positively with cardiovascular risk,28 unstable angina and poorer outcome of coronary artery disease.29

In contrast, adiponectin has direct antiatherogenic and anti-inflammatory actions. It inhibits monocyte adhesion to endothelial cell walls (through inhibiting NF-κB signalling, thus suppressing TNFα-induced adhesion molecule expression), suppresses the transformation of macrophages into foam cells30,,31 and reduces proliferation and migration of smooth muscle cells.18 In addition, variation in the ADIPOQ gene promoter, which regulates the synthesis of adiponectin, may be directly related to intima-media thickness of the carotid artery.32

Adipokines and insulin resistance

When discussing systemic insulin resistance it is important to consider the insulin-sensitive tissues and organs, such as adipose tissue, the liver and, importantly, skeletal muscle, which is the body's largest glucose metabolizer. Increased release of NEFAs and adipokines by excess adipose tissue leads to insulin resistance by disrupting the insulin signalling cascade. TNF-α acts locally to directly interfere with insulin signal transduction in adipocytes by down-regulating several steps in the insulin signalling cascade. It increases serine phosphorylation of insulin receptor substrate-1 (IRS-1), which inhibits insulin receptor tyrosine kinase activity.33 This is sufficient to block the downstream events of insulin signalling, including expression and translocation to the cell membrane of the insulin-sensitive glucose transporter GLUT4, thus reducing insulin-induced glucose uptake. In skeletal muscle, TNF-α also impairs insulin sensitivity by reducing the activity of IRS-1 and downregulating GLUT4.34,35

TNF-α also stimulates the expression of other adipokines, such as leptin, PAI-1 and IL-6, and increases the release of NEFAs, which may contribute to insulin resistance in other tissues.36 NEFAs inhibit insulin-stimulated glucose metabolism in muscle and stimulate gluconeogenesis in the liver.37 IL-6 interferes with insulin signalling by increasing the expression of the suppressor of cytokine signalling family of proteins that are involved in the degradation of IRS.38

Resistin has been reported to induce insulin resistance in skeletal muscle, liver and adipose tissue; although its function remains unknown it is postulated to be involved in regulation of metabolism, or regulation of adipogenesis, or to have a role in inflammation.39

Insulin resistance is associated with low levels of adiponectin.40 The protective effects of adiponectin are mediated in several ways. In the liver, adiponectin increases insulin sensitivity by lowering NEFA uptake, increasing fatty acid oxidation and reducing hepatic gluconeogenesis and very low-density lipoprotein (VLDL) production. In muscle, adiponectin stimulates glucose uptake and fatty acid oxidation.40

Cholesteryl ester transfer protein and dyslipidaemia

Adipose tissue is a prominent source of CETP, and the activity and mass of CETP is increased in obesity.21 CETP transfers cholesteryl esters from HDL-C to VLDL and LDL, and triglycerides from VLDL to LDL and HDL (Figure 2).41 When the activity of CETP is high and the level of VLDL particles is increased (resulting from increased NEFA influx into the liver) HDL cholesteryl esters are preferentially transferred to VLDL, increasing the cholesterol content and making them more atherogenic. In addition, the HDL particles become smaller and denser, and are cleared from the circulation more rapidly. CETP also interacts with hepatic lipases that promote the formation of small, dense LDL particles.

Figure 2.

Mechanism of CETP. Adapted from Expert Rev Cardiovasc Ther 2004; 2: 485–501 with the permission of Future Drugs Ltd.

The net effect of increased CETP activity is the pattern of dyslipidaemia often referred to as the ‘atherogenic lipid triad’ seen in patients with metabolic syndrome or type 2 diabetes—increased triglyceride levels, decreased HDL-C levels and a greater proportion of small, dense LDL. This pattern of dyslipidaemia is particularly atherogenic—in particular small, dense LDL-C particles have a greater propensity to form oxidized LDL and are less rapidly cleared from the circulation.42


Obesity and type 2 diabetes are both associated with an increased likelihood of developing NAFLD. This condition is associated with insulin resistance and a systemic inflammatory state indicated by increased levels of CRP.43 While the pathophysiology of NAFLD has not been completely elucidated, it is thought to stem from the accumulation of triglyceride in hepatocytes in combination with oxidative stress, lipid peroxidation, proinflammatory cytokines (e.g. TNF-α, IL-6) and adipokines.7

Insulin resistance in the liver, possibly contributed to by TNF-α produced by adipose tissue, may be the underlying cause that leads to NAFLD. Adiponectin and leptin may also have a direct role in the development of NAFLD.7 Adiponectin increases fatty acid oxidation and reduces fatty acid synthesis in hepatocytes—reduced levels of adiponectin in obese people, therefore, remove this protection against lipid accumulation.44 Adiponectin also suppresses hepatic TNF-α production. Leptin deactivates IRS in hepatocytes leading to insulin resistance, a clinical feature commonly seen in patients found to have steatosis on ultrasound together with raised liver enzymes such as alanine aminotransferase.

Targeting adipokines to reduce cardiometabolic risk

The impact of weight loss

Effective lifestyle modification to achieve weight loss can have a significant impact on the risk of developing type 2 diabetes. In the Finnish Diabetes Prevention Study, the risk of developing diabetes was reduced by 58% over about 3 years in patients with impaired glucose tolerance, who received individualized counselling regarding diet and exercise to achieve their weight-loss and dietary goals.45 A similar risk reduction was seen in the US Diabetes Prevention Programme, with the incidence of type 2 diabetes also being reduced by 58% in patients assigned intensive lifestyle modification.46 Furthermore, in this study lifestyle intervention was significantly more effective than treatment with metformin at reducing the incidence of type 2 diabetes (5% vs. 7.8% for each year in the study).

Improving diet and increasing physical exercise in order to reduce weight produce changes in adipokine levels. In a study involving 120 obese women followed for 2 years, subjects who received education about how to lose weight through a Mediterranean-style diet and increased exercise had greater weight loss than those who received general information about healthy food and exercise (–14 kg vs. –3 kg, respectively, P < 0.001).47 Serum levels of IL-6 and CRP were significantly lower and adiponectin significantly higher in the intervention groups compared with controls, suggesting that greater weight loss was associated with greater improvements in these markers and was the result of diet selection and education.47

Lifestyle changes, such as improved diet and increased exercise, can be difficult to initiate or adhere to for many people and so pharmacological therapy to promote weight loss is an alternative approach. Orlistat and sibutramine are licensed for the treatment of obese patients or overweight patients who present with significant co-morbidities, such as type 2 diabetes or dyslipidaemia.48,49

Orlistat was one of the first licensed pharmacotherapies shown to reduce body weight by an additional 2.9 kg above placebo in clinical trials in conjunction with a weight reducing diet at 12 months.50 This is accompanied by a modest decrease in blood pressure (2 mmHg), blood glucose, LDL-C and triglyceride levels, but no significant effects were seen in HDL-C levels.50 In addition, orlistat is recommended three times a day and is associated with side effects such as fatty/oily stool, faecal urgency and oily spotting, which occurred in 15–30% of subjects in clinical trials.

Compared to placebo, sibutramine achieved an additional weight loss of 4.3 kg in patients treated for 1 year.50 While significant increases in HDL-C and substantial decreases in triglycerides were seen with sibutramine in the STORM trial,51 these results have not been consistent across its clinical trials.52 Sibutramine has little effect on LDL-C, and in some patients increases particularly diastolic blood pressure and heart rate. For these reasons the drug is not recommended in patients with uncontrolled hypertension, pre-existing CVD or tachycardia, pathologies not uncommon in patients with obesity.50 Sibrutamine is also known to interact with drugs that inhibit CYP3A4 such as ketoconazole and erythromycin.

Studies investigating the effects of fat removal by liposuction and lipodectomy on metabolic parameters have yielded conflicting results. Improved insulin sensitivity and lower levels of circulating inflammatory markers have been reported following fat removal in obese patients.53 However, another study reported no effects on insulin sensitivity or inflammatory markers such as CRP, IL-6, TNFα and adiponectin.54 This could be because these procedures remove mainly subcutaneous adipose tissue rather than visceral adipose tissue, which has more direct access to vasculature.

Therapies with anti-inflammatory properties

Abnormal lipid levels have been shown to be important risk factors for CVD.55 Pharmacological therapies used to modify lipid levels, such as statins, niacin and fibrates, have also demonstrated anti-inflammatory properties. Statins inhibit 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, a key enzyme in the biosynthesis of cholesterol, and are used to lower LDL-C levels and reduce the risk of vascular events.56 They have also been shown to have anti-inflammatory effects in addition to their lipid-lowering properties, as evidenced by reductions in CRP levels independent of LDL-C lowering.57 This anti-inflammatory effect may contribute to reductions in cardiovascular risk seen with statin treatment. In the PROVE-IT trial, patients treated with statins who achieved a CRP level below 2 mg/l had better clinical outcomes than those who did not regardless of the LDL-C level achieved.58

Niacin lowers LDL-C and triglyceride levels and increases HDL-C. Nicotinamide, a major metabolite of niacin, has been shown to inhibit several cytokines, including IL-6 and TNF-α.59 In addition, niacin appears to have an anti-inflammatory effect, as it reduces CRP plasma levels in a dose-dependent manner 60 and short-term treatment with extended-release niacin has been shown to increase adiponectin levels.61 In patients with diabetes, while lower doses of extended-release niacin are effective, high doses of niacin can impair glycaemic control.60

The peroxisome proliferator-activated receptors (PPARs) are lipid-activated transcription factors that regulate genes involved in lipid and glucose metabolism. PPARα is expressed mainly in the liver, skeletal muscle and kidney and is involved in fatty acid oxidation. The fibrates, such as bezafibrate and fenofibrate, are PPARα agonists and are used to increase HDL-C and reduce triglyceride levels. While the results of the FIELD study were inconclusive,62 other studies have shown fibrates to reduce fatal and non-fatal MI and CHD.63,64 PPARα is also expressed in endothelial cells, smooth muscle cells and monocytes/macrophages, and so PPARα agonists also interfere with monocyte recruitment and adhesion.65 This therapeutic strategy is promising, therefore.

PPARγ is abundantly expressed in adipocytes and its signalling pathways are involved in the control of lipid uptake, transport, storage and metabolism. The glitazones are PPARγ agonists that improve glucose metabolism and insulin action in patients with type 2 diabetes. In the PROACTIVE trial, pioglitazone significantly reduced the risk of death, non-fatal MI and stroke in patients with type 2 diabetes and signs of macrovascular disease.66 However, the difference between groups for the composite primary endpoint was not significant and treatment with pioglitazone was associated with higher rates of heart failure than placebo (11% vs. 8%, P < 0.0001).

Two small studies investigating the effects of glitazones on lipoproteins, inflammatory markers and adipokines in non-diabetic patients with metabolic syndrome showed both pioglitazone and rosiglitazone to significantly reduce CRP and resistin levels while increasing adiponectin concentrations.67,68 In addition, pioglitazone significantly increased HDL-C levels and reduced the concentration of small, dense LDL-C, and rosiglitazone significantly reduced IL-6. It is worth noting that glitazones promote weight gain despite improving insulin resistance, although this is associated with redistribution of weight from visceral to subcutaneous depots.69 More recently, some adverse attention has been aimed at glitazones as a cause of increased rates of cardiovascular events, a claim which is being challenged.70,71

The anti-inflammatory effects of aspirin (acetylsalicylic acid) are achieved through inhibiting prostaglandin production by cyclooxygenase enzymes. Prostaglandins are also involved in blood clotting, and low-dose aspirin is used to reduce the risk of thrombosis that can cause MI or stroke. In a meta-analysis of 25 trials, aspirin reduced vascular mortality by ∼15% and MI or stroke by around 15% in patients who had previously experienced a transient ischaemic attack, stroke, unstable angina or MI.72 The benefits of aspirin in the primary prevention of cardiovascular events were first demonstrated in the Physicians’ Health study73 and the British Doctors’ Trial.74 In addition, a meta-analysis of six major trials of aspirin in the primary prevention of CVD showed that aspirin reduced the risk of CHD, non-fatal MI and total cardiovascular events, although there were no significant differences in the incidences of stroke or cardiovascular mortality.75 Aspirin has also been shown to reduce CRP and IL-6 in patients with stable angina.76 Aspirin is amongst the most effective methods of reducing cardiovascular events but at higher dose is associated with increased bleeding risks.77

There is increasing evidence to suggest that angiotensin II, the key effector of the renin–angiotensin system, is capable of inducing an inflammatory response in the vasculature wall. Blockade of the renin–angiotensin system can be achieved with angiotensin converting enzyme (ACE) inhibitors and angiotensin II receptor blockers. Both classes of antihypertensive medication have been reported to improve endothelial function and reduce atherosclerosis-associated events.23 Angiotensin II receptor blockers and ACE inhibitors appear to have beneficial effects beyond their ability to reduce blood pressure; for instance, the ACE inhibitor ramipril significantly reduced death, MI and stroke in patients at high risk of cardiovascular events.78

Targeting the endocannabinoid system

The endocannabinoid system (ECS) plays a role in the regulation of energy intake.79 Stimulation of the ECS induces food intake and the ECS may become hyperactive in response to a high-fat diet as endogenous endocannabinoids are synthesized from arachidonic acid, which is in turn derived from essential fatty acids obtained in the diet. Sustained hyperactivity of the ECS may, therefore, contribute to the development of obesity and related cardiometabolic risk factors.79 The selective cannabinoid receptor 1 antagonist, rimonabant, blocks activation of the ECS thus reducing food intake and facilitating weight loss. The efficacy of rimonabant has been shown in four large, placebo-controlled randomized studies, the Rimonabant in Obesity trials, in which patients receiving rimonabant achieved weight loss of 3.9–5.4 kg greater than placebo while the drug was being administered.80–83 While patient tolerance was high for rimonabant, psychiatric disorders (depression, anxiety, irritability, aggression) were 3% more likely to occur in patients treated with rimonabant compared with placebo.84

In addition to weight loss, rimonabant displayed beneficial effects on several other cardiometabolic risk factors. In overweight and obese patients with type 2 diabetes on monotherapy, rimonabant reduced HbA1c by 0.7% more than placebo (P < 0.0001), an effect only partly explained by the observed weight loss alone.83 Rimonabant increased HDL-C levels and reduced triglyceride levels by 8.1 and 12.4%, respectively, more than placebo (P < 0.001 for both).82 While overall levels of LDL-C did not change significantly, the distribution of LDL particle size moved towards a lower proportion of small-dense LDL.82 In two of the Rimonabant in Obesity trials, the effects of rimonabant on HDL-C and triglyceride levels were greater than could be attributed to weight-loss alone.80,83 Rimonabant also significantly increased adiponectin levels, an effect that was partly independent of weight loss. This weight loss-independent effect has also been shown in vitro, where addition of rimonabant to adipocyte cultures produced a rapid increase in adiponectin levels.85 Rimonabant is the third agent licensed as an adjunct to diet and exercise for the treatment of obese adults (body mass index ⩾30 kg/m2) or overweight adults (body mass index >27 kg/m2) with associated risk factor(s) such as type 2 diabetes or dyslipidaemia. Recently, the National Institute for Health and Clinical Excellence (NICE) published its Final Appraisal Determination (FAD), which recommended that rimonabant be made available as an adjunct to diet and exercise to overweight or obese patients who have not responded to, are intolerant of, or are contraindicated to sibutramine and orlistat.86

While several agents have been licensed directly for weight loss, data on long-term safety, tolerability and more importantly cost effectiveness is still being compiled.87 Moreover, these current agents are effective only while being taken, with weight regain on discontinuation. Newer more potent agents with the ability to help sustain lifestyle modifications even after discontinuation would be more ideal targets of the future.


The production of adipokines by excess adipose tissue contributes to inflammation, dyslipidaemia, insulin resistance, endothelial dysfunction and a prothrombotic state, all of which increase the risk of CVD. Weight loss, either by improving diet and exercise or pharmacological means, can produce benefits in terms of decreased adipokine levels as adipose mass decreases. In addition, pharmacological therapies can reduce adipokine levels in a weight loss-independent manner; for example, through anti-inflammatory effects or direct action on adipocytes. Orlistat and sibutramine both achieve weight loss and the novel cannabinoid receptor 1 antagonist rimonabant achieves weight loss in addition to modest improvements in several other cardiometabolic risk factors simultaneously. All agents have some limitations and weight regain on discontinuation, however, the potential of cardiovascular risk reduction provided above and beyond weight reduction warrants wider use of appropriate agents in economies where these agents are shown to be cost effective.

Conflict of interest: AB is a consultant, has received honoraria as a speaker and recruits patients to trials funded by pharmaceutical companies including Astra Zeneca, Sanofi-Aventis, Merck, Amenarini, Roche, Takeda, Abbott and Health-Smart.


View Abstract