Medical management of the patient considering bariatric surgery

2.1. Hypothyroidism 

Hypothyroidism is associated with modest weight gain despite a reduction in appetite, in large part due to fluid retention. It is also associated with a constellation of other symptoms and signs of insidious onset including lethargy, constipation, cold-intolerance, hair loss, brittle nails, dry skin, menorrhagia, husky voice (due to myxoedema of the vocal cords and oropharyngeal structures) and peri-orbital puffiness. The full-blown picture of myxoedema with thickened features, non-pitting oedema, enlarged tongue and extreme mental and physical slowing is now, thankfully, rare due to the readily available and reliable assays for thyroid stimulating hormone (TSH) and free thyroid hormones. In practice, it is extremely rare for a patient to be referred to specialist services without thyroid function tests being performed, as this represents a simple and readily available screening test, which may be performed by the general practitioner. The request for a blood test may also frequently be made by the patient convinced that “their glands” are the cause of their obesity. The majority of obese patients will be found to have normal thyroid function. In the rare instances that frank hypothyroidism is detected; levothyroxine therapy is clearly indicated to normalise TSH and free thyroid hormones. Free thyroid hormones will normalise rapidly within days of starting medication but the elevated TSH generally takes longer to normalise, over a number of weeks. However, by far the commonest abnormality of thyroid function is sub-clinical hypothyroidism i.e. elevated TSH in the presence of normal free thyroid hormones. This is a common condition occurring in 3–8% of the general population. It is more common in women than men, and its prevalence increases with age.13 The most important implication is high likelihood of progression to clinical hypothyroidism, particularly if thyroid auto-antibodies are positive. It remains contentious whether sub-clinical hypothyroidism is an independent risk factor for cardiovascular disease and large-scale randomized studies to provide an evidence-base for levothyroxine therapy are lacking. Small studies support treatment with levothyroxine in patients with TSH of more than 10.0mIU/L, whereas in patients with TSH levels lower than this, quality of life scores are not improved but anxiety scores are significantly increased.14 In fact 80% of patients with sub-clinical hypothyroidism have a serum TSH of less than 10mIU/L. The current pragmatic approach is routine levothyroxine therapy for patients with a persistent serum TSH greater than 10.0mIU/L, as they have greatest risk of progression to frank hypothyroidism. For those with a TSH of less than 10.0mIU/L, therapy should be individualized and based on other factors including co-morbidities, family history of thyroid disease, desire for pregnancy and thyroid auto-antibody levels.13 It seems reasonable, therefore, to initiate a trial of levothyroxine therapy to reduce TSH levels to the lower end of the normal range in the obese subject as part of medical management or pre-surgical optimisation. As the normal range is wide for free thyroid hormones, normalisation of TSH into the lower end of the normal range is the best guide to adequate replacement. This should be achieved pre-operatively in the small proportion of patients who have thyroid dysfunction pre-operatively, however, in practice, this does not result in the dramatic weight loss that patients may be hoping for.

2.2. Cushing's syndrome 

Central obesity is associated with type 2 diabetes mellitus, hypertension and dyslipidaemia. This cluster of risk factors, known as the metabolic syndrome, also occurs in people with primary glucocorticoid excess or Cushing's syndrome (CS). Although plasma glucocorticoid levels are not elevated in obesity, recent research suggests that this clinical similarity may reflect enhanced tissue glucocorticoid levels due to increased activity of the enzyme 11beta-hydroxysteroid dehydrogenase type 1 (11beta-HSD1).15 Given the high and increasing prevalence of central obesity and the relatively low prevalence of CS, this then begs the question, who should we be testing and which investigation should we be using? It would be easy to advocate universal screening, as for hypothyroidism, if an inexpensive, highly sensitive and specific screening test were available. However, in practice, biochemical differentiation between obese subjects with and without CS can be extremely difficult. Obese patients not only share clinical features with CS patients, but also co-morbidities commonly associated with obesity, such as depression and OSA, may lead to false positive screening tests for Cushing's syndrome. Unlike screening for thyroid disease, which is performed by a simple test with no restriction on timing of the sample, screening for Cushing's syndrome is extremely complex and should be performed by an appropriately experienced specialist endocrinologist or following a local protocol agreed with an appropriately experienced endocrinologist. Timing of samples and, where appropriate, administration of dexamethasone, requires meticulous attention to detail as does the exclusion of interfering medications. Meaningful interpretation of the results requires accurate documentation of all of the above as well as clinical assessment of prior probability of Cushing's syndrome.

A recent international consensus on the diagnosis of CS evaluated three widely used first-line screening tests: 24-h urinary free cortisol (UFC), low-dose dexamethasone suppression tests (LDDSTs) and late-night salivary cortisol.16

Measurement of UFC is widely used as a screening test for CS. A 24-h UFC collection provides an integrated measure of serum cortisol, smoothing out diurnal variations in plasma cortisol, and is unaffected by factors influencing cortisol binding globulin (CBG) levels. UFC values more than 3–4-fold above the upper limit of normal are virtually diagnostic of CS.17 Conversely, normal cortisol excretion in three 24-h samples makes CS unlikely.18 The sensitivity and specificity of this test have been reported to range from 95 to 100%, and 98%, respectively, although others have found much lower sensitivity and specificity.18 However, to achieve reasonable sensitivity and specificity the test requires three complete 24-h urine collections, and does not represent a quick and easy screening test. Furthermore, UFC levels may be normal in mild or cyclical CS, or with incomplete collections. Conversely, elevated UFC levels may be found in 40–60% of patients with depression (a common association of obesity), in polycystic ovarian syndrome (PCOS) and in pregnant women.19

The overnight dexamethasone suppression test (DST) involves the oral administration of 0.5–2mg of dexamethasone at 23.00–24.00h and measurement of serum cortisol at 08.00h or 09.00h the next morning. The 1mg DST has been advocated by many as the screening test of choice for diagnosis of CS.13 This test has been reported to be generally reliable in the obese patient, but with a slightly higher false positive rate.20

Using a diagnostic cut-off of 50nmol/l plasma cortisol (with sample taken at 09.00h), the sensitivity of the overnight DST probably approaches 100%. Certainly, adequate suppression on this test effectively rules out a diagnosis of CS unless the index of suspicion is very high and cyclical Cushing's syndrome is suspected. However, the specificity has been reported as relatively low. False positives probably occur more often in obese subjects, but a recent report in 100 obese individuals has reported a reduction in the false positive rate from 8% to 2% when using a dose of 2mg rather than 1mg.21 Pragmatically, a 2mg overnight DST seems a reasonable screening test, with any subjects failing this test requiring further investigation with a 2-day LDDST. All of these tests require obsessional attention to detail in terms of timing of dexamethasone dosing, sample collection, exclusion of any factors affecting dexamethasone metabolism and interpretation of final results. As such, they are best carried-out by specialist endocrine units with experience in diagnosis and management of CS. Certain clinical features have a higher positive predictive value for CS than others, notably thin skin, easy bruising, pigmented striae and proximal myopathy. However, as we evaluate more patients with obesity, some early, mild cases of CS are identified which often lack the classical clinical features. Although universal screening may not be practically possible, any approach using targeted screening will inevitably miss the occasional case of mild early hypercortisolaemia. A strong case can certainly be made for screening in all those with coexisting diabetes or hypertension.

2.3. Other acquired (endocrine) causes of obesity 

Other rare endocrine disorders can include obesity as part of their presentation. Insulinoma, a rare insulin producing islet cell tumour, is associated with symptomatic hypoglycaemia and usually with significant weight gain as sufferers learn to eat regular snacks to avoid symptoms. Hypoglycaemia is most typically worst after an overnight fast or skipping a meal, but can occur post-prandially and may be associated with episodes of loss of consciousness. Endocrine referral for evaluation with supervised fast and possibly prolonged oral glucose tolerance test is indicated. Acquired lesions of the hypothalamus including craniopharyngioma, langerhans histiocytosis and sarcoidosis can present with weight gain due to disruption of hypothalamic neuro-circuits involved in regulation of energy balance. Other features of hypothalamic disease may be present including polyuria and polydypsia due to diabetes insipidus, labile blood pressure and disordered thermoregulation.

2.4. Monogenic obesity syndromes 

A number of single gene defects have been described which result in obesity, including mutations in leptin and its receptor, proopiomelanocortin (POMC) and the melanocortin 4 receptor (MC4R).6, 22 Most of these conditions are very rare, exhibit autosomal recessive inheritance and are associated with extreme obesity and hyperphagia from early childhood. MC4R mutations represent the commonest single gene defect causing obesity, as heterozygous loss of function mutations lead to a phenotype of older onset obesity and have been identified in up to 8% of series of morbidly obese adults. A more extreme phenotype with early onset obesity is seen in patients with homozygous or compound heterozygous mutations. These mutations are associated with disordered hypothalamic regulation of appetite and energy balance.

Monogenic obesity should be considered where there is onset of obesity very early (in the pre-school years) associated with marked hyperphagia. In the case of leptin deficiency, treatment with subcutaneous leptin reverses the obesity syndrome and associated endocrine effects, but so far only a handful of families with this condition have been identified worldwide. POMC mutations are associated with red hair (due to impaired melanin production) and adrenal dysfunction. Prader-Willi syndrome (PWS) is associated with early onset obesity with hyperphagia and extremely persistent food seeking behaviours. The associated features of neonatal hypotonia, short stature due to growth hormone insufficiency and learning and behavioural difficulties mean that this condition is more likely to have been diagnosed in childhood, but some cases may not be detected until later in life. Except for leptin deficiency, specific therapies are not available for these monogenic conditions. Management of PWS is centred on behavioural therapy and hormone replacement where indicated.

Sufficient data is lacking on the effects of bariatric surgery on inherited syndromes of appetite dysregulation and acquired hypothalamic disease. However, a less favourable outcome would be expected following gastric bypass surgery as the associated changes in gut hormones and reduction in appetite may be less effective if central responses to these peripheral signals are impaired.

3.1. Diet/lifestyle 

Patients who have been referred for bariatric surgery should already have been offered dietary and lifestyle intervention and a therapeutic trial of anti-obesity drugs. However, in practice the degree of lifestyle support that has been offered is variable. Diet – balanced, low calorie restriction – remains the cornerstone of management but often has poor long-term effect due to a complex interaction of psychological, metabolic and genetic factors. Similarly, sustainable exercise programmes resulting in negative energy balance have produced long-term effects, although weight loss is limited to about 5%.

Despite variable response rates, lifestyle modification is worthwhile as responders experience significant reduction in co-morbidities. The Lifestyle Heart Trial24 confirmed a reduction of angina and an improvement in coronary angiographic changes with a 12kg weight loss over one year in those compliant with a very low fat diet and regular yoga. Several studies have confirmed benefits of weight loss in25, 26 preventing/delaying progression to diabetes in those at high risk.

3.2. Pharmacotherapy 

Drug therapies may have been offered in a “fire and forget” style of prescription, without adequate warning of potential side effects, resulting in poor patient compliance. An example of this is the occurrence of steatorrhoea after consumption of high-fat foods whilst taking orlistat. One of the roles of the physician in the obesity MDT is to ensure that patients have received an adequate (minimum 6 months) trial of medical or drug therapy, as recommended by NICE. This may involve dose adjustment to ameloriate any side effects and ensuring appropriate duration of therapy to assess response. Currently available weight-reducing drugs include orlistat and sibutramine.

Orlistat, a gastrointestinal lipase inhibitor, reduces dietary fat absorption and so reduces energy intake. It is prescribed as a 120mg tds regimen. Orlistat is indicated for those with BMI>30 or BMI>28 with existing co-morbidity e.g. diabetes, hypertension, dyslipidaemia. It has beneficial effects on LDL-cholesterol greater than that expected from weight loss alone. It also reduces blood pressure, HbA1c in diabetes, serum triglycerides and waist circumference. The XENDOS trial of over 3000 individuals with a BMI>30 confirmed a 37% reduction in the risk of developing diabetes in those treated with Orlistat 27, 28 over the 4 year trial period.

Sibutramine is a centrally acting satiety inducer, which acts by inhibiting reuptake of serotonin and noradrenaline. This should be initiated at the 10mg dose and increased to 15mg according to clinical response (if <2kg weight loss at 4 weeks). Due to its pharmacology, there are several contraindications including co-administration of MAOIs, pyschiatric illness and uncontrolled hypertension. The principal adverse effects are an increase in BP and heart rate, sweaty episodes, and insomnia. Fortnightly BP checks are advised to monitor any BP rise, particularly in the initiation phase.

Both drugs have been reported to assist in a greater than 10% weight loss in 12–18% of the treated groups in clinical trials lasting over one year. In the same trials, between 21% and 32% achieved a greater than 5%27, 28 weight loss compared to placebo.

4.1. Diabetes 

The judicious use of carefully selected diabetes pharmacotherapy, in combination with lifestyle and pharmaceutical weight loss interventions, may prevent the need for bariatric surgery in some patients and reduce the risk by improving glycaemic control and pre-operative weight in others. Many therapies to control hyperglycaemia have weight gain as an adverse effect. This has been observed with oral sulphonylureas (e.g. glibenclamide, gliclazide) and glitazones (rosiglitazone, pioglitazone) and also with subcutaneous insulin with either a basal or multiple injection regimen. Choosing alternative therapies or lower doses may ameliorate this side effect. The basal analogues (insulin glargine, insulin detemir) compared to isophane insulins appear to have less weight gain and are the preferred choice if these insulins are required in management. There is some evidence that insulin detemir compared to insulin glargine, has even less impact on weight gain in some patients.29

By contrast, the newer incretin-based diabetes treatments have been associated with significant weight loss. The injectable glucagon-like peptide-1 (GLP-1) agonist, exenatide (BYETTA trade name), is the first available incretin mimetic, although daily and weekly incretin mimetics are being developed. As described above, GLP-1 is released from the gut and inhibits food intake by reducing appetite and also slowing gut emptying. GLP-1 is also a powerful incretin, promoting glucose-dependent insulin secretion in response to a meal and reducing glucagon secretion from pancreatic alpha cells. The combined actions to reduce appetite and gastric emptying and to promote insulin secretion in response to a meal, make this a useful treatment for type 2 diabetes. Byetta has nausea as a side effect which may be minimised by administering the subcutaneous injection closer to meal times and is often reduced after the initiation phase. In some individuals, significant weight reductions may occur with improvement of glucose control. It is given as twice a day subcutaneous injection. It is approved by NICE for patients with type 2 diabetes with inadequate blood glucose control (HbA1c≥7.5%) after a trial of metformin and sulfonylurea, and with a BMI>35kg/m2, in those of European descent (adjusted as appropriate for other ethnic groups) or in those with “BMI<35 in whom insulin is unacceptable because of occupational implications or where weight loss would benefit other co-morbidities”.

In practice, a number of centres are using exenatide in patients with lower BMIs in an attempt to avoid severe obesity in patients struggling with weight gain. Pancreatitis has been reported in patients taking exenatide, although rates are very low. The incidence of pancreatitis is increased in patients with type 2 diabetes compared to the background population and a causal relationship has not been established. However, abdominal pain in any patient taking exenatide should lead to prompt exclusion of pancreatitis.

The oral dipeptidyl peptidase-4 (DPP-4) inhibitors such as sitagliptin or vildagliptin, also act on the incretin system, prolonging the half-life of endogenous GLP-1 by inhibiting breakdown by the enzyme DPP-4. Like metformin, they are usually weight neutral. Use of these agents may allow reduction or30, 31 withdrawal of other oral hyperglycaemic agents with a less favourable weight profile.

Particularly critical is the management of glycaemic control in the peri-operative phase. Often, increasing doses of insulin are required to improve glucose control before surgery with a down-titration in the post-operative phase.

4.2. Hypertension 

Increasing body weight is associated with a rise in BP and the development of hypertension. Sympathetic overactivity is increased in overweight individuals and this contributes to vasoconstriction and potentially, renin production. Additional mechanisms linking obesity and hypertension are endothelial dysfunction and arterial stiffness. Obesity is also linked to other cardiovascular risk factors (e.g. impaired glucose tolerance and dyslipidaemia) that amplify cardiovascular risk. Weight reduction per se, (even if still overweight) confers benefits in BP reduction as well as improving the other cardiovascular risk factors.32 Blood pressure (BP) measurement in the overweight patient groups requires close attention to BP cuff size to avoid measurement error.33 Using a BP cuff too small for the arm circumference leads to a significant over estimation of BP. The bladder of the BP cuff needs to encircle at least two-thirds of the upper arm circumference. Most commercially available bladders for the BP cuff are 23–35cm long. The alternative or large BP cuff bladder is 42cm long and should be used in muscular or obese individuals. In marked obesity often there is a need to use a “thigh” BP cuff.

In the treatment of hypertension, all classes of antihypertensives may be effective. However there are distinct adverse drug effects which may need to be considered in the management of obesity-associated hypertension. Several groups of drugs have unwanted effects in overweight or obese patients. Both beta-blockers and thiazide diuretics may worsen glucose tolerance and dyslipidaemia. Worsening of exercise tolerance and weight gain has been observed with beta-blockers. This class of drugs should be introduced at low-dose only if there appears to be resistant hypertension. Indapamide which has minimal impact on glucose or lipid metabolism is a useful alternative to thiazide diuretics. Calcium blockers are associated with ankle oedema and best used at standard doses to avoid this effect. Angiotensin Converting Enzyme Inhibitors (ACEIs) may produce problematic cough in a few patients, with either angiotensin 2 blockers (losartan, irbesartan, candesartan, valsartan or olmesartan) or direct renin inhibitors (alleskirin) being alternatives.

Maximising BP control in the peri-operative period may require combination drug therapies using any of the antihypertensive drug classes in higher than usual doses. Following successful weight loss with surgery a full review of all drug therapy needs to be undertaken which should allow appropriate down-titration of drugs and/or dose.