The Clinical Biochemistry of Obesity

Abstract

Obesity is essentially an excessive accumulation of triacylglycerols in fatty tissue that is the net result of excessive energy intake compared to energy usage. Severe forms of the disease are most likely to have a predominantly genetic basis and this is probably polygenic. The ‘thrifty gene’ hypothesis also describes the disturbance that a modern environment, including higher energy intake and decreased physical activity, has on otherwise advantageous genetic variations. While the physical consequences of obesity, such as arthritis, are debilitating and costly, the metabolic consequences are the drivers behind the modern epidemics of insulin resistance, diabetes, fatty liver disease, coronary artery disease, hypertension and polycystic ovary syndrome. The pathophysiological mechanisms behind these diseases are probably a combination of the toxic metabolic effects of free fatty acids and adipokines - the numerous messengers that adipose tissue has been discovered to produce.

Introduction

Obesity is a condition of increased adipose tissue mass.¹ Obesity can also be defined as an increase in body weight beyond the limits of physical requirement, as the result of an excessive accumulation of fat. Accumulation of fat, or triacylglycerol, is essentially the only way that body weight can become excessive, as other energy storage (e.g. carbohydrate glycogen or protein in liver and muscle) does not have the potential of adipose tissue to exceed the limits of requirement. Although anabolic steroids can increase lean body mass and therefore body mass, this has only been described in those already malnourished.²

Adipose tissue is a tissue entity that can, through hyperplasia and hypertrophy, vary enormously between individuals, more so than any other tissue. However it is misleading to think of it as a single entity, as there are subtypes of adipose tissue (e.g. visceral and subcutaneous) which appear to have different implications for health.³

Adipose tissue is not purely a storage tissue for triacylglycerols, it acts as an endocrine organ also,^4,5 releasing numerous chemical messengers (adipokines) that communicate and affect other tissues.

This review considers the changes in clinical biochemistry measurements that are associated with obesity, and the insights into the pathophysiology behind this most important health issue in western and developing countries.

Definitions

The definition of obesity cannot be simply made in terms of body weight because we should expect short people to be lighter than tall people. Therefore we need to standardise body weight against body height. The simplest expression for this is the body mass index (BMI) calculated as weight (kg) divided by height squared (m²). The critical importance of this weight for height adjustment is illustrated in its origins from life insurance tables.⁶ A simple prediction of life risk was interpreted as a weight that was 20% above the average for frame size, which was equivalent to a BMI of 27.8 (kg/ m²).⁷ The World Health Organisation (WHO) guidelines of 1985 defined obesity as a BMI >30.0 for men and >28.6 for women.⁸ Although women have lower bone and muscle mass, they usually have slightly more subcutaneous fat but these subtleties are often ignored in standardised approaches. Similarly, both muscle mass and bone mass decrease with age and so like sex, age should be considered as a variable of interest in many obesity studies and standards. The definitions were further refined by the WHO with a BMI over 25 being defined as ‘overweight’ and over 30 as being ‘obese’.⁹ Finally, there are racial differences in body composition that should also be taken into account.¹⁰

It should be also stated that the ratio of waist to hip circumference (normally below 0.95 in men and 0.85 in women) is generally a better prognostic indicator for disease than BMI^11–13 especially when the BMI is less than 35, and there may be advantages in applying both measurements.¹⁴

Acquired Causes of Obesity

The prevalence of overweight and obesity varies from country to country but in Western countries like Australia, it is becoming true that most of the population can be affected.¹⁵ The prevalence is also increasing in children.¹⁶ The variation from country to country and from time to time implies that environmental factors are the major determinant of disease prevalence. While obesity is thought to be the second most preventable cause of death behind smoking, a recent study suggests that the health care costs of obesity exceed those of smoking.¹⁷

Fundamentally, obesity is the result of excessive energy intake compared to energy expenditure. In children, increased energy intake as sugar¹⁸ or fat¹⁹ has been linked to obesity, as has decreased physical activity in children.²⁰ However even in children this is not a simple problem as other factors such as low weight in infancy can also predict later obesity.²¹

Cushing’s syndrome may cause obesity. It is also associated with truncal or visceral obesity, which can be difficult to differentiate from simple obesity. This distinction is one of the main purposes of tests such as low dose dexamethasone suppression tests used to differentiate Cushing’s syndrome from obesity.

Although slight decreases in energy expenditure in clinical or subclinical hypothyroidism may contribute to weight gain,²² hypothyroidism is a rare cause of obesity and much of the weight gain is due to water retention which is reversible after thyroid hormone treatment. Insulinoma can cause massive weight gain due to the excessive energy intake consumed to avoid hypoglycaemia but is an extremely rare condition and therefore a very uncommon cause of obesity.

Normally signals from the gut and adipose tissue are integrated in the central nervous system to affect appetite and energy homeostasis and limit weight gain. Pathological obesity may result from the failure of these homeostatic mechanisms²³ although our understanding of these processes is still relatively rudimentary.

Genetic Causes of Obesity

The idea that some people are born with a tendency to obesity is not new and Hippocrates stated that “sudden death is more common in those who are naturally fat than in the lean”. But why would nature allow such genes to exist? The basic premise of the ‘thrifty gene’ hypothesis^24,25 is that certain populations may have genes that determine increased fat storage, which in times of famine represent a survival advantage, but in a modern environment result in obesity and type 2 diabetes.²⁶ Identification of such thrifty gene candidates may help provide insight into the pathogenetic processes of the numerous physical inactivity-mediated disorders.²⁷

Underweight newborns become overweight children who become overweight adults^28,29 although this natural progression is being questioned.³⁰ There is a suggestion that homeostatic set points for adipose tissue mass and insulin sensitivity may be set both by genetic factors and by energy metabolism in utero (Barker hypothesis). ³¹

Twin studies have shown important (up to 75%) genetic explanation to BMI.^32,33 With the exception of the rare mutations that cause severe morbid obesity, it seems that numerous genes, each with modest effect, contribute to an individual’s predisposition toward the more common forms of obesity.³⁴ Some genetic syndromes often have obesity as part of a larger syndrome of manifestations and these include Prader-Willi, Angelman and Wilson-Turner syndromes.

Genome-wide scans for obesity susceptibility genes have been performed in several populations of diverse ethnic backgrounds, and many have been replicated in corresponding studies.³⁵ Although there have been hundreds of loci with high log of the odds scores, some of the most promising include 1p36 (D1S468 a tumour necrosis factor alpha (TNFα) receptor gene), 2q14 (D2S410 a gene associated with high triglyceride levels) and 6q27 (a locus associated with transient neonatal diabetes mellitus).³⁶ In contrast, candidate gene approaches look for mutations in genes that are presumed to be relevant.

Leptin Associated Genes

Leptin is secreted from adipocytes into the circulation, traverses into the central nervous system and binds to leptin receptors in the hypothalamic arcuate nucleus. This stimulates the production of pro-opiomelanocortin (POMC). The two products of POMC are alpha-melanocyte stimulating hormone (alpha-MSH) and adrenocorticotropin (ACTH). Alpha-MSH binds to melanocortin-4 receptors in the hypothalamic paraventricular nucleus which cause a decrease in food intake.

It appears that a genetically lean individual will gain an extra 7 to 8 kg before leptin increases sufficiently to stop weight gain. Individuals who gain more must be unresponsive to the hormone either because it cannot enter the brain sufficiently or because there is a mutation in one of the many steps of leptin action.³⁷ Mutations of this system (leptin,³⁸ leptin receptor,³⁹ POMC,⁴⁰ alpha MSH receptor⁴¹) are generally uncommon or rare but can cause obesity. Although it was hoped that leptin deficiency caused obesity, we now know that obesity usually has high leptin levels and leptin resistance is likely.⁴²

Beta-3 Adrenergic Receptor (ADRB3) Gene

This is expressed in adipose tissue and is involved in the regulation of lipid metabolism and thermogenesis. A missense mutation (Trp64Arg) has a high frequency in Pima Indians⁴³ and has been frequently associated with obesity in other populations.^44–46 The interaction of this receptor with other receptors may also affect its ability to couple with its mediators such as G-proteins.

Peroxisome Proliferator Activated Peptide Receptor Gamma (PPAR-γ) Gene

The importance of this receptor in obesity is supported by the efficacy of thiazolidinediones in the treatment of visceral obesity. These drugs, also known as glitazones, bind avidly to the PPAR-γ receptor⁴⁷ leading to improved insulin sensitivity with major changes in fat metabolism including a reduction in free fatty acids (FFA)⁴⁸ by improved peripheral and subcutaneous uptake.⁴⁹ They also decrease insulin levels presumably by reducing FFA, reducing the release of TNF-α and restoring adiponectin levels.⁵⁰ Note that this is not the same receptor as PPAR-α which also potentiates FFA catabolism in the liver but is the molecular target of the lipid-lowering fibrates (e.g. gemfibrozil).⁵¹

PPAR-γ is a nuclear receptor that is important for adipogenesis and insulin signalling. The Pro12Ala mutation is common and results in a decreased ability to bind to PPAR-γ responsive genes. The mutation has effects on BMI that are variable but it could be that its greatest effect is on individuals that are already predisposed to obesity.⁵² Similarly individuals with the ADRB3 Trp64Arg mutation are far more likely to be obese if they also have the Pro12Ala PPAR-γ mutation.⁵³ Other genes associated with the PPAR-γ, including its coactivator-1 (PGC-1), have also been found to have many alleles associated with obesity.⁵⁴

Adiponectin Gene

This adipocyte derived peptide has had many regulatory actions on energy homeostasis, glucose and lipid metabolism and anti-inflammatory pathways described. High levels of adiponectin generally lead to weight loss. Polymorphisms of the adiponectin gene have also been associated with obesity and insulin sensitivity.⁵⁵

Physical Pathology of Obesity

Osteoarthritis is one of the major costs of obesity. Osteoarthritis in the knees and ankles may be directly related to the trauma associated with the degree of excess body weight.⁵⁶ Non-weight bearing joints may still be affected by altered cartilage and bone metabolism.

Obstructive sleep apnoea is a physically defined entity characterised by the absence of airflow in the presence of thoracoabdominal movements. 70% of patients with obstructive sleep apnoea are obese⁵⁷ and this may be due to neck fat and fat deposits in the pharyngeal area.⁵⁸ Decreases in residual lung volume are associated with increased abdominal pressure on the diaphragm.⁵⁹ There is variable lowering of nocturnal oxygen saturation, which is usually mild, and measurement of oxygen saturation has limited diagnostic use.⁶⁰ When underlying pulmonary disease is absent, only major degrees of obesity affect pulmonary function. While only 5% of all obese patients have obstructive sleep apnoea, almost half have loud snoring and a third have excess daytime sleepiness. Obstructive sleep apnoea has been associated with all of the diseases of obesity including hypertension and coronary artery disease (CAD). Pulmonary hypertension is also possible due to vasoconstriction of pulmonary arterial bed during apnoea that extends into the waking hours. But although obstructive sleep apnoea has been suggested as a cause of pulmonary hypertension, it is not recognised as a risk factor on its own.

Finally one of the consequences of obesity is community stigma where public disapproval may affect education, employment, income, marital status and health care. These are significant detrimental effects on the quality of life and are associated with higher incidence of depression.

Obesity and Insulin Resistance

The risk of diabetes increases by 9% for each kg gained in self reported weight⁶¹ and generally starts to increase at a BMI of 22⁶² and is 40 times higher at a BMI over 35.^63,64

Insulin resistance is widely recognized as a fundamental defect seen in obesity and type 2 diabetes. The development of type 2 diabetes is strongly associated with overweight and obesity in both genders and all ethnic groups. Over 90% of diabetics are overweight or obese.⁶⁵ Weight gain and insulin resistance usually precede the onset of diabetes. Current theories indicate that type 2 diabetes develops when pancreatic beta cell output can no longer satisfy the demands imposed by increased insulin resistance.⁶⁶

The FFA Paradigm Linking Obesity to Insulin Resistance

The predominant paradigm used to explain insulin resistance is the elevated FFA concentrations in visceral obesity. The importance of a portal source of FFA and its direct access to the function of the liver could explain the insulin resistance of the liver with central obesity.⁶⁷ Increased adipose tissue stores, a disturbed insulin-mediated regulation of lipolysis and subnormal skeletal muscle FFA uptake under conditions of high lipolytic rate may further increase circulating FFA concentrations.⁶⁸ In addition, a disturbance of FFA uptake by adipose tissue post-prandially is also a critical determinant of plasma FFA concentration.

Elevated cellular levels of FFA can produce insulin resistance in skeletal muscle and liver, as well as reduce beta-cell function, and this has been referred to as lipotoxicity.⁶⁹ Several lines of evidence indicate that hepatic FFA and triglyceride accumulation are a causative factor involved in hepatic insulin resistance.⁷⁰ Evidence is increasing that insulin-resistant muscle is characterised by a lowered ability to oxidise FFA. Perturbations in FFA metabolism occur during accumulation of skeletal muscle triglyceride and may also be implicated in the pathogenesis of insulin resistance.⁷¹ An imbalance between FFA uptake and FFA oxidation may in turn be a factor promoting accumulation of lipid intermediates and triacylglycerols within skeletal muscle, which is strongly associated with skeletal muscle insulin resistance.⁷² FFA can block insulin-signalling pathways and lead to insulin resistance.⁷³ Unsaturated FFA may serve as a nutrient sensor to determine whether FFA are to be stored or oxidized and thereby reduce the risk of developing fatty liver and insulin resistance.⁷⁴ Chronically elevated FFA contribute to beta cell dysfunction by significantly increasing the basal rate of insulin secretion.^75,76 However it is also believed that the beta-cell changes found in diabetes are better correlated with increased glucose levels than with FFA levels, thus supporting an importance of glucotoxicity.⁷⁷

The Adipokine Paradigm Linking Obesity to Insulin Resistance

This paradigm focuses on adipose tissue as an endocrine organ.⁷⁸ Recent studies have suggested that adipokines (adipose tissue-derived hormones and inflammatory cytokines) play essential roles in overall insulin sensitivity and the dysfunctions of adipose tissue which can lead to systemic insulin resistance.⁷⁹ This concept is not independent of FFA theories as there still seems to be a relationship between FFA levels and adipokines particularly in individuals who are overweight or insulin resistant.⁸⁰

Cytokines

Cytokines produced by visceral adipose tissue are thought to be of possible major importance with the most studied of these adipose cytokines being TNF-α.⁸¹ TNF-α is produced by adipose tissue⁸² and its expression is elevated in the adipose tissue in multiple experimental models of obesity. TNF-α inhibits the synthesis of several other adipocyte-specific proteins including adiponectin and enhances the release of FFA from adipose tissue.⁸³ Neutralisation of TNF-α in one of these models improves insulin sensitivity by increasing the activity of the insulin receptor tyrosine kinase, specifically in muscle and fat tissues. On a cellular level, TNF-α is a potent inhibitor of the insulin-stimulated tyrosine phosphorylation on the beta-chain of the insulin receptor and insulin receptor substrate-1, suggesting a defect at or near the tyrosine kinase activity of the insulin receptor. Given the clear link between obesity, insulin resistance, and diabetes, these results strongly suggest that TNF-α may play a crucial role in the systemic insulin resistance of NIDDM. TNF-α can stimulate IL-6, which, in turn, stimulates the acute phase reactant production of CRP, Plasminogen activator inhibitor 1 (PAI-1) and fibrinogen from the hepatocyte.⁸⁴ Fibronectin is also elevated and shows some correlation with insulin, but not C-peptide or measures of body weight.⁸⁵

Leptin

Leptin was discovered in 1994. The ‘ob/ob’ mouse lacks the ability to produce or respond to leptin resulting in severe obesity.⁸⁶ Leptin decreases neuropeptide Y in the hypothalamus and should suppress appetite. Fat mass is the primary determinant of serum leptin in humans with energy intake and gender also having significant effects.⁸⁷ Gender influences leptin production and reactivity,⁸⁸ presumably through the reproductive hormones.⁸⁹ Catecholamines also influence leptin production and the leptin signal to the appetite centre.⁹⁰ Additional regulators of leptin production include glucocorticoids, cytokines and agonists of PPAR-γ. Leptin is not only produced by adipose tissue but is also produced in several other places including placenta, bone marrow, stomach, muscle and perhaps brain, thus increasing the number of potential regulatory roles for this hormone.⁹¹

Adiponectin

Adiponectin is a novel adipose tissue-specific protein that has structural homology to collagen VIII and X and complement factor C1q, and circulates in human plasma at high levels.⁹² Adiponectin expression and/or secretion is increased by insulin like growth factor-1 and decreased by glucocorticoids and beta-adrenergic agonists. Adiponectin expression and secretion is increased by activators of PPAR-γ.⁹³ Adiponectin exhibits potent anti-inflammatory and anti-atherosclerotic effects⁹⁴ including inhibiting the expression of TNF-α induced endothelial adhesion molecules, macrophage-to-foam cell transformation, TNF-α expression in macrophages and adipose tissues, and smooth muscle cell proliferation.⁹⁵ Production is reduced in insulin resistance indicating that the degree of hypo-adiponectinaemia is more closely related to the degree of insulin resistance and hyperinsulinaemia than to the degree of adiposity or glucose intolerance.⁹⁶ Adiponectin’s effects seem to be peripherally mediated and the evidence of an association between adiponectin and the metabolic and cardiovascular complications of obesity is growing all the time.⁹⁷

PAI-1

PAI-1 is the primary physiological inhibitor of plasminogen activation in blood and is known to contribute to thrombus formation and to the development and the clinical course of acute and chronic cardiovascular diseases. Plasma levels of PAI-1 are regulated on a genetic basis but, more importantly visceral fat accumulation is considered as a major regulator of PAI-1. Expression by adipose tissue could be responsible for the elevated plasma PAI-1 level observed in insulin resistance. While adiponectin has anti-atherogenic properties, it is also inversely related to PAI-1⁹⁸ which is closely involved in the development of atherosclerosis. Elevated PAI-1 level is a core feature of insulin-resistance⁹⁹ and pro-inflammatory cytokines may have an important role in PAI-1 over-expression.¹⁰⁰ It is suggested that PAI-1 may not merely increase in response to obesity and insulin resistance, but may have a direct causal role in obesity and insulin resistance.¹⁰¹

Adipsin

Adipsin is a serine protease that is secreted by adipocytes and belongs to the alternative complement pathway (complement D).¹⁰² It is deficient in mouse models of obesity however this may be a secondary phenomenon.¹⁰³

Resistin

Resistin is an adipokine with putative pro-diabetogenic properties.¹⁰⁴ Although there is evidence that circulating levels are proportional to the degree of adiposity, levels are not related to the degree of insulin resistance.^105,106

Metabolic Syndrome Criteria

The metabolic syndrome (previously known as syndrome X) has insulin resistance as its hallmark as indicated in the WHO classification of metabolic syndrome¹⁰⁷ (Table 1). The third report of The National Cholesterol Education Program (NCEP) Expert Panel also developed criteria¹⁰⁸ (Table 2) which are similar, but can lead to differences in classification of various populations.^109,110

Table 1.

WHO Criteria for Metabolic Syndrome: Insulin resistance (Hyperinsulinaemia and/or Fasting Glucose >=6.1) + 2 of the following factors:

	MEN	WOMEN
Body Mass Index	>= 30 kg/m2	>= 30 kg/m2
Or Waist Hip Ratio	>0.9	>0.85
Triglycerides	>1.7 mmol/L	>1.7mmol/L
HDL Cholesterol	<0.9 mmol/L	<1.0 mmol/L
Microalbuminuria	>2.5mg/mmol creatinine	>2.5mg/mmol creatinine
Blood Pressure	>=140/90 mmHg	>=140/90 mmHg

Table 2.

NCEP Criteria for Metabolic Syndrome: 3 of the following factors:

	MEN	WOMEN
Waist Circumference	>= 102cm	>=88cm
Triglycerides	>=1.7 mmol/L	>=1.7mmol/L
HDL Cholesterol	<=1.0 mmol/L	<=1.3 mmol/L
Fasting Glucose	>=6.1 mmol/L	>=6.1 mmol/L
Blood Pressure	>=130/85 mmHg	>=130/85 mmHg

Although hyperuricaemia is also related to insulin resistance and was included in the original syndrome, the relation between serum urate and the risk of coronary heart disease depends heavily upon the presence of pre-existing myocardial infarction and widespread underlying atherosclerosis as well as the clustering of risk factors.^111,112

Advanced Tests of Insulin Resistance

The so-called ‘gold-standard’ test of insulin resistance is the euglycaemic clamp which requires the infusion of glucose and insulin, and is therefore only useful for intensive physiological studies on small numbers of subjects. Furthermore, caution should be exercised when making comparisons between studies due to variations in infusion protocols, sampling procedures and hormone assays used. The minimal model approach is a frequently sampled IV glucose tolerance test but is also best suited to a research setting as it still requires up to 30 blood samples.

Simple Tests of Insulin Resistance

The simplest test is a fasting or random glucose level, however this is insensitive particularly as we do not know how much insulin is being secreted to maintain that glucose level.

The next simplest test is an insulin level. The main problem here is that insulin levels are highly variable from minute to minute, let alone after meals. This is no longer due to insulin assay imprecision but due to the pulsatile release of insulin coupled with its short half-life. A single fasting level is still so variable that it may easily be misleading and many suggest that at least three fasting levels should be taken and averaged to obtain a better estimate of the usual fasting insulin level.

Simple estimates of insulin sensitivity and pancreatic beta-cell function using fasting insulin and glucose levels serve as surrogate measures of insulin sensitivity and secretion.¹¹³ A very simple tool is the fasting insulin to glucose ratio. A level greater than 4.5 (using SI units) has been described as being useful in the diagnosis of insulin resistance polycystic ovary syndrome¹¹⁴ and greater than 7.0 in girls with premature adrenarche.¹¹⁵

An alternative calculation based on fasting insulin and glucose is the ‘Homeostatic Model Assessment’ (HOMA), as it is assumed that the fasting state is homeostatic. This calculation is essentially the product of fasting insulin and glucose concentrations and is more useful than either measure on its own. Note that it will also be influenced by the variability of single fasting insulin estimates as well as the insulin assay chosen. The HOMA calculation can be configured to be a measure of resistance (HOMA-R) or sensitivity (HOMA-S). It is more appropriate for large epidemiological studies however it is important to be aware that advanced tests of insulin resistance measure stimulated insulin resistance whereas HOMA gives an estimate of basal insulin resistance. Normal HOMA-R levels are awkward to define across age groups.¹¹⁶

The quantitative insulin sensitivity check index (QUICKI),¹¹⁷ was designed to give a more linear relationship with clamp estimates than the HOMA (which is inversely and reciprocally related to clamp values). Both QUICKI and inverse of HOMA-R are useful measures¹¹⁸ and suitable for diagnosis of insulin resistance in clinical and epidemiological practice¹¹⁹ and only fail in unusual clinical scenarios.¹²⁰ However, a normal QUICKI reference interval needs to be established for each laboratory with an appropriate control group because of significant inter-laboratory variations in insulin values as well as differences in various populations.¹²¹

My own view is that most of the variation of both the HOMA and QUICKI correlates to fasting insulin variation and therefore up to 90% of the information can be obtained from the fasting insulin level alone. Insulin levels can also be measured during an oral glucose tolerance test and I have found them practically useful in assessing normoglycaemic individuals at high risk of insulin resistance. No single test of insulin resistance will be appropriate under all circumstances¹²² and the tests should not be assumed to give equivalent assessments.^123,124

Clinical Associations of Obesity

Liver Disease

Fat accumulation in the liver is independent of body mass index, intra-abdominal and overall obesity but characterized by several features of insulin resistance in normal weight and moderately overweight subjects.¹²⁵ Increased hepatic VLDL production is associated with insulin resistance and the high rate of triglyceride turnover is often greater than the ability to secrete. Insulin resistance may also result in an inability to suppress apo B degradation.¹²⁶ Hepatic steatosis (fatty liver) consists of small or large intracytoplasmic lipid droplets especially around terminal hepatic veins (zone 3).¹²⁷

Impaired triglyceride export and an insufficient increase in free FFA mitochondrial beta oxidation could aggravate the situation leading to the presence of oxidisable lipids in hepatocytes and could also trigger lipid peroxidation, mitochondrial dysfunction and cytokine production.¹²⁸ Non-alcoholic steatohepatitis (NASH) is a combination of steatosis with necro-inflammatory changes including enlarged hepatocytes, apoptotic bodies, Mallory bodies and giant mitochondria with loss of cristae. Peripheral insulin resistance, increased FFA beta-oxidation, and hepatic oxidative stress are present in both fatty liver and NASH, but NASH alone is associated with mitochondrial structural defects.¹²⁹ Inflammation appears with lymphocytic and neutrophilic infiltrates usually around altered hepatocytes or in the portal areas.

Fibrosis and cirrhosis may occur around the terminal hepatic veins and then form bridges between terminal hepatic veins or between adjacent portal tracts. Fatty liver and NASH is increasingly being recognised as an important cause of liver related morbidity and mortality¹³⁰ and is believed by many to be one of the most common causes of cryptogenic cirrhosis.¹³¹ The morbidly obese can be expected to have fatty liver changes including portal inflammation and fibrosis in 30% and cirrhosis in 3%.¹³²

Although ferritin levels have also been found to be predictive of fatty liver¹³³ it is important to recognise that ferritin levels will be increased when ALT levels are elevated and may be secondary to fatty liver rather than related to cause.

Laboratory abnormalities in fatty liver include a 2 to 4-fold elevation of serum transaminase levels with other liver function test results usually normal.¹³⁴ NASH is becoming the most common reason for referral for investigation of abnormal liver function tests. Central adiposity, hyperleptinaemia, and hyperinsulinaemia were the major determinants of the association of overweight with elevated serum ALT activity.¹³⁵ Performing oral glucose tolerance testing in cases with fatty liver disease may be useful for early screening of diabetes mellitus.¹³⁶

Body weight, rather than alcohol consumption, may be the major factor in determining the serum level of liver enzymes. Even when body weight is not generally considered to be overweight, slight to moderate gains in weight are associated with increases in serum liver enzymes.¹³⁷ Laboratories should determine age-adjusted reference intervals for enzymes in children, and gender-adjusted reference intervals for transaminases, gamma-glutamyltransferase, and total bilirubin in adults.¹³⁸ These reference intervals for ALT may include variations due to BMI¹³⁹ with body weight explaining 12% of the normal variation of ALT.¹⁴⁰

The clinical adage “fat, female, fertile and forty” indicates that gallstone incidence is higher in the overweight and in fact very high in obesity.¹⁴¹ Cholesterol production increases as body fat increases¹⁴² (10 kg body weight gain is equivalent to an extra egg a day) and high concentrations of cholesterol relative to bile acids will increase the likelihood of gallstone precipitation.

Coronary Artery Disease

BMI increases the risk of CAD, and weight gain from any initial BMI further increases the risk (especially weight gain of 20 kg or more).

Dyslipidaemia may be the most important relationship of BMI to CAD.¹⁴³ Obesity increases VLDL (triglycerides) through increased production and decreased clearance of triglyceride rich lipoproteins due to lack of stimulation of lipoprotein lipase.¹⁴⁴ Obesity also lowers HDL in men and women of all ages^145–147 and ethnicities.¹⁴⁸ While LDL levels are not consistently elevated in obesity, LDL is smaller and denser¹⁴⁹ and more atherogenic.

Cholesterol ester transport protein (CETP) exchanges triglycerides from VLDL to LDL in exchange for cholesterol esters. This results in triglyceride rich LDL particles that are rapidly lipolysed by hepatic lipase leaving smaller denser LDL particles. Small dense LDL can more easily be oxidised or glycated possibly leading to less identification by the LDL receptor and decreased clearance. Possibly small dense LDL is also more likely to get through endothelial fenestrations.

CETP also exchanges triglycerides from VLDL to HDL in exchange for cholesterol esters. This similarly results in triglyceride rich HDL particles that are rapidly lipolysed by hepatic lipase allowing HDL to be cleared from the circulation.

High fasting triglyceride levels (i.e. VLDL) predict the presence of small dense LDL in diabetes,^150–152 non-diabetics¹⁵³ and hypopituitarism.¹⁵⁴ Remarkably the value of triglyceride that predicts the small dense LDL phenotype is about 1.5 mmol/L, the same level used in the definitions of metabolic syndrome. More specifically the logarithm of the triglyceride concentration is inversely related to particle size.¹⁵⁵ The total cholesterol to HDLC ratio is more predictive than the LDLC to HDLC ratio because it mathematically includes also the component Trig/HDL which is crucially abnormal in insulin resistance and predicts the presence of small dense LDL.

Hypertension

Hypertension is present in about half of all overweight individuals¹⁵⁶ and obesity alone accounts for about 70% of essential hypertension.¹⁵⁷ Cardiac weight increases with increasing body weight, but heart weight as a percentage of total body weight is lower than in normal weight controls. In obesity, the increase in cardiac output is not explained by the presence of the new adipose tissue, but may be due to increased sympathetic activity. Adrenaline (from the adrenal medulla) tends to be normal to low in obesity and there is a decreased response of adrenaline to hypoglycaemia and exercise.¹⁵⁸ However, noradrenaline levels (from sympathetic nerve endings) tend to be higher.¹⁵⁹ Hypertension in the overweight is associated with increased sympathetic activity^160,161 and sympathetic blockers have greater effect in obesity.¹⁶² The causes of sympathetic overactivity include hyperinsulinaemia, increased intrarenal pressures, hepatic FFA, angiotensin II, leptin, central chemoreceptor sensitivity and impaired baroreceptor reflex.¹⁶³

Sodium reabsorption is increased with high fat diets¹⁶⁴ and the renin/angiotensin/aldosterone system (RAAS) is activated in obesity despite volume expansion and sodium retention.¹⁶⁵ Aldosterone tends to be higher in obese individuals while renin is often relatively normal and there is a positive correlation between BMI and the aldosterone to renin ratio.¹⁶⁶ Adipose has long been expected to produce a factor that directly affects RAAS and recently all the components of the renin-angiotensin system have been found to be fully represented in the adipose tissue. Furthermore, they appear to be up-regulated in obesity and circulating angiotensinogen levels are enhanced¹⁶⁷ due to elevated adipose angiotensinogen gene expression in obesity.^168–170

Other findings of interest include the decreased levels of atrial and ventricular natriuretic peptides (natural antagonists of the RAAS) which may also help to explain the susceptibility of the obese to hypertensive disorders.¹⁷¹

Renal hyper-filtration together with glucose intolerance, hyperlipidaemia and hypertension can lead to obesity related focal segmental glomerulosclerosis.

Polycystic Ovary Syndrome (PCOS)

PCOS has been described as ‘the thief of womanhood’ as it is commonly associated with oligomenorrhoea and hirsutism.¹⁷² Multiple ovarian cysts are actually a common ultrasound finding (up to 20% in 18–25 y/o) however this finding is usually not associated with infertility, although it may be associated with hirsutism.¹⁷³ It is generally the association of obesity with multiple ovarian cysts that leads to infertility.¹⁷⁴

A universally accepted definition of PCOS does not exist, however most modern definitions acknowledge the association of insulin resistance together with hyperandrogenism and infertility. Up to 30% of all PCOS women have impaired glucose tolerance and an additional 7.5% have diabetes while even 10.3% of non-obese women with PCOS have impaired glucose tolerance and 1.5% have diabetes.¹⁷⁵ 16% of PCOS women develop diabetes by the age of menopause.¹⁷⁶ Conversely, up to 27% of pre-menopausal women with type 2 diabetes will also have PCOS.¹⁷⁷

Sex hormone binding globulin (SHBG) is usually low in PCOS. The most important hormone that SHBG binds is testosterone. The presence of increased total testosterone in PCOS is uncommon compared to the prevalence of increased free testosterone estimates.¹⁷⁸ Free testosterone is believed to be the biologically active form but is difficult to measure, however it is clear that if there is little binding protein, more testosterone must be free and active. SHBG levels can be increased by oestrogen or thyroxine,¹⁷⁹ however low levels are usually due to androgens or insulin.¹⁸⁰ Levels of SHBG do not correlate with androgen levels but rather with body mass index^181,182 and insulin resistance.^183,184 SHBG is predictive of NIDDM in women^185,186 and similarly predictive of overall mortality in post-menopausal women.¹⁸⁷ Abnormal lipid levels are seen in 70% of women with PCOS particularly if they are obese.¹⁸⁸

Increased LH pulse frequency and amplitude occurs in PCOS. This may be in part due to the effects of elevated free testosterone.¹⁸⁹ The normal LH/FSH ratio is below 2.0, whereas in PCOS it usually rises to over 2.5. Marked elevations of both LH and FSH are seen normally in mid-cycle when this ratio is less discriminatory.

Other Endocrine Effects of Obesity

Obesity has no effect on TSH, FT4, TRH, thyroglobulin or total T4. Reverse T3 (rT3) has been negatively correlated to BMI but it is already known that overfeeding can lead to increased levels of T3 (where rT3 is decreased) and this is the most likely relationship.¹⁹⁰

Growth hormone (GH) levels are lower in obesity.¹⁹¹ The circadian rhythm is maintained but GH responsiveness is also diminished and there are fewer GH pulses with lower amplitude.¹⁹² Chronic hyperinsulinaemia in obesity may stimulate IGF-1 production but simultaneously suppress hepatic IGFBP-1 and IGFBP-2 production, which may result in inhibition of IGF-1 bioactivity.¹⁹³ Changes in IGF-1 levels may also be responsible for the lower GH levels, through negative feedback.¹⁹⁴

Cortisol production is increased, but as there is also increased metabolism, basal levels of cortisol (and ACTH) are normal in obesity. Abnormalities in the pituitary adrenal axis have been described. However we need to distinguish between simple obesity and the obesity associated with Cushing’s syndrome, otherwise it is difficult to assess the importance of differing responses in cortisol and ACTH stimulatory tests in obesity.¹⁹⁵

Fat distribution develops in adolescence and androgens and oestrogens produced by the gonads and adrenals as well as the peripheral conversion of androstenedione to estrone in fat cells are pivotal in body fat distribution. Oestradiol and oestrone levels are increased in obese men probably due to the increased peripheral conversion of testosterone and androstenedione to estradiol and estrone. Despite the high levels there is generally no evidence of feminisation.

Total testosterone levels are lower in obese men, largely due to decreases in SHBG.¹⁹⁶ However, free testosterone levels seem to be normal and libido, testicular size, potency and spermatogenesis are also usually normal. Obesity has been proven to affect other aspects of sexual function as it is independently associated with erectile dysfunction and may improve with weight loss.¹⁹⁷ LH levels are usually normal¹⁹⁸ however gross obesity may cause low LH levels.¹⁹⁹ Low free testosterone levels may occur²⁰⁰ as evidenced in sleep apnoea syndrome,²⁰¹ and this may be mediated by the significantly increased oestrogens in obese men.

Finally vitamin D levels are also lower in obesity^202,203 leading to higher PTH levels.²⁰⁴ Social isolation and an indoor existence may be significant factors.

Conclusion

The rate of obesity is increasing throughout the world. Environmental changes continue to occur in developed and developing countries creating a global pandemic with enormous implication of morbidity and mortality in the coming decades. It seems that each of us will have a polygenetic risk to developing obesity and can only hope that the improving understanding of the causes and complex relationships of obesity will lead to better prevention and treatments. The clinical biochemistry laboratory often provides useful evidence for clinicians and patients when the overt signs of the condition are either hidden or denied. Until then we only need to look around us for motivation in our personal battles with this disease!