Controversy still surrounds the classification of acute and chronic pancreatitis. The most recent classifications define acute pancreatitis as a spectrum of inflammatory lesions in the pancreas and peripancreatic tissue with variable degrees of oedema, necrosis, and haemorrhage. Exocrine and endocrine function may be impaired to a variable extent and for a variable length of time. If the cause can be rectified and complications such as pseudocysts dealt with, then a clinical return to functional and morphological normality is usual. Acute pancreatitis rarely progresses to chronic pancreatitis.

Chronic pancreatitis is almost always associated with recurring or persisting pain. Chronic inflammation causes fibrosis with destruction of exocrine parenchyma leading to malabsorption and steatorrhoea. In the later stages, diabetes mellitus may follow destruction of the endocrine parenchyma.

Chronic pancreatitis is relatively rare. Autopsy estimates of prevalence range from 0.03 to 0.4 per cent, but care has to be taken in the interpretation of postmortem findings. Pancreatic calculi are not uncommon after the age of 70 and are often associated with atrophy and fibrosis. However, these changes may be the result of squamous metaplasia of the ductal epithelium, are not associated with alcohol abuse, and do not produce the clinical picture of chronic pancreatitis.

In Europe and North America the incidence of chronic pancreatitis appears to be increasing, although this may reflect increased awareness and improved diagnosis. The prospective Copenhagen Pancreatitis Study found an annual incidence of 8.2 cases per 100000 and a prevalence of 27.4 per 100000. Extrapolation from these figures predicts a mean life expectancy of only 3.3 years from diagnosis, an estimate which is clearly too low, and it may be that the prevalence figure is an underestimate. In general, countries with a low alcohol intake have a low incidence of chronic pancreatitis, although there are many discrepancies. For example, Switzerland has a higher per capita alcohol consumption than Denmark and Sweden but a much lower incidence of chronic pancreatitis.

Two forms of chronic pancreatitis are generally recognized; a common ‘calcific’ form and a less common ‘obstructive’ form. Some accept the existence of a third ‘inflammatory’ form which can only be identified on histological examination and will not be considered further.

This type of pancreatitis is found in association with alcohol abuse (which now accounts for some 70 per cent of all cases in Western centres), hypercalcaemia, and malnutrition, and in hereditary and idiopathic chronic pancreatitis. The disease has a patchy lobular distribution, at least in its early stages, and inflammation leads to fibrosis, destruction of parenchyma, and eventual atrophy. Dilatation of the pancreatic duct system is often prominent. A common pathogenesis has been proposed in which the ductal system becomes occluded by protein precipitates or plugs which subsequently calcify. It is postulated that overstimulation of the acinar cells deranges intracellular transport of secretory proteins, resulting in an abnormal admixture of digestive enzymes and lysosomal hydrolases and/or storage of zymogens in acid compartments. Secretion of the iron binding protein, lactoferrin, is also increased; this molecule is known to associate strongly with acidic molecules to form complexes, a property that might facilitate formation of protein precipitates. In addition to secretory protein, pancreatic juice also contains a so-called pancreatic stone protein which can prevent nucleation and growth of calcium carbonate crystals.

Alcohol raises the viscosity of pancreatic juice by increasing secretory protein output and lactoferrin concentration, but it may also decrease biosynthesis of pancreatic stone protein (or increase its denaturation). Thus, the stage is set for the formation of protein plugs in the lumen of the acini and small ducts, and for their subsequent calcification. The calculi so formed consist of calcium carbonate, usually in the form of calcite (which constitutes approximately 95 per cent of their weight) and eosinophilic protein fibrils, which are thought to be a degraded form of pancreatic stone protein.

Although alcohol is heavily implicated in protein plug formation, less than 10 per cent of alcoholics develop chronic pancreatitis, and other factors must be involved. In addition, many patients with chronic pancreatitis have not abused alcohol and some are teetotal. Evidence suggesting that alcohol is more likely to cause chronic pancreatitis in individuals consuming a diet high in fat and protein is now contested: inadequate intake of antioxidant vitamins and trace elements may be more important. Induction of the detoxifying P450-I cytochromes occurs in the majority of patients with chronic pancreatitis regardless of cause. Such induction increases the production of oxygen free radicals and reactive toxic intermediates derived from ingested drugs and chemicals. If antioxidant intake is insufficient, failure of normal protective mechanisms could allow organelles within pancreatic acinar cells to become damaged. The trace elements zinc, copper, and manganese are required for activity of the free radical scavenger superoxide dismutase, while selenium is an essential component of glutathione peroxidase, an enzyme which controls endogenous peroxidase formation from unsaturated fatty acids in cell membranes. Alcoholics who smoke are at a higher risk of developing chronic pancreatitis than those who do not: cigarette smoking may increase free radical formation while at the same time diminishing antioxidant levels and the ability to scavenge free radicals.

A tropical form of chronic calcific pancreatitis is well recognized in underdeveloped areas of Africa and India, with a particularly high incidence in the Indian state of Kerala. Tropical pancreatitis characteristically presents with recurrent abdominal pain in childhood, development of diabetes at around puberty and death in early adulthood. The pancreatitis is associated with marked calcification and was once thought to result from the protein calorie malnutrition of kwashiorkor. However, it now seems more likely that pancreatic pain and insufficiency cause malnutrition rather than result from it. Dietary toxins, such as the cyanogenic glycosides found in cassava, a staple consumed in many areas affected by tropical pancreatitis, have also been implicated.

The incidence of chronic pancreatitis in patients with hyperparathyroidism has fallen from 5 to 10 per cent to around 1 to 2 per cent, probably as a result of earlier diagnosis and treatment of the parathyroid disorder. Hypercalcaemia stimulates the acinar cell and may also increase the concentration of calcium in pancreatic juice, thus favouring calculi formation.

Hereditary pancreatitis is a rare autosomal dominant disorder with incomplete penetrance. Pancreatitis is of early onset, calcification is marked, and there may be hyperlipidaemia and disorders of amino acid metabolism. The condition may increase the risk of pancreatic cancer, a risk reported to be as high as 25 per cent in some series. As in idiopathic pancreatitis (which is now thought to account for 10 to 30 per cent of cases of chronic pancreatitis) the mechanisms involved in the production of pancreatic inflammation are obscure but reduction in pancreatic stone protein levels may be involved.

Obstruction of the duct due to papillary stenosis, scarring from acute pancreatitis or trauma, cysts or pseudocysts, or neoplasia may produce generalized diffuse inflammation and duct dilatation in the pancreas upstream of the point of blockage. Recurrent attacks of acute inflammation frequently precede chronic inflammation and more persistent pain. Characteristically, these acute attacks are less severe clinically than those which often develop in acute pancreatitis due to gallstones, and life threatening progression to necrosis and abscess formation is exceptional.

It is debatable whether pancreas divisum is a cause of obstructive pancreatitis. In pancreas divisum there is failure of fusion of the embryological dorsal and ventral pancreas so that most of the exocrine secretion has to pass through the smaller accessory papilla to gain the duodenum. The widespread use of endoscopic retrograde cholangiopancreatography (ERCP) indicates that pancreas divisum may be present in up to 10 per cent of the population, calling into question its role in pancreatitis. Nevertheless, pancreas divisum may be present in some patients with chronic pancreatitis and it may be such patients have superimposed papillary stenosis.

In contrast to chronic calcific pancreatitis, there may be considerable clinical and morphological improvement in obstructive pancreatitis once the obstructing lesion has been dealt with.

Only about 5 per cent of patients with chronic pancreatitis do not experience pain. In the remainder pain is usually the cardinal symptom, and it is by far the most common indication for surgery. The pain is recurrent or persistent, often incapacitating, and frequently requires administration of opiates for relief. It is usually experienced in the upper abdomen with radiation through to the back and is poorly localized, but may be lateralized to one or other hypochondrium and lower chest. The back pain can be particularly distressing. Some patients gain relief by leaning forwards while sitting, while some resort to kneeling on all fours. Lying flat in bed can be extremely painful and many patients sleep with the most painful side uppermost. Pain is often, but by no means invariably, exacerbated by food and alcohol intake and fatty foods may be particularly troublesome. Patients frequently apply heat to painful areas and permanent skin mottling (erythema ab igne) may result.

Attacks of acute inflammation may cause exacerbations of pain lasting for some days in the early stages of chronic pancreatitis. Although such exacerbations seldom produce a severe clinical illness, each attack should still be managed from the outset as a potentially life-threatening illness. With progression of chronic pancreatitis, acute exacerbations become less frequent and less severe, and serum amylase and lipase elevations become less marked as the gland loses its secretory capacity.

Obstruction of the pancreatic duct may cause attacks of pain which begin after eating and last for some hours. Pressure in the pancreatic duct is increased in most patients with chronic pancreatitis and it is tempting to assume that eating triggers exocrine secretion, increasing duct pressure and causing duct dilatation and pain. Pancreatic interstitial pressure is also raised and in general, operations designed to improve drainage reduce ductal and interstitial pressure. However, some patients with duct dilatation do not appear to suffer much pain, while others have pain in the absence of dilatation.

Pancreatic cysts and pseudocysts are common in patients with chronic pancreatitis and may cause pain as a consequence of the pressure within the fluid collection. While drainage often relieves pain, there is a variable relationship between pain, the presence of pseudocysts, and the magnitude of intracyst pressure.

Attention has centred recently on morphological changes in the nerves supplying the pancreas and their role in pain production. The mean diameter of the nerves is increased while the mean area served by each nerve is reduced. Oedema in nerve bundles is common and the perineurial sheath appears to serve as a less effective barrier between the surrounding connective tissue and its contents. Perineurial inflammation may be present, although its extent does not correlate with pain severity. Immunohistochemical studies suggest that there are increased amounts of neurotransmitters, such as substance P, in the nerve sheaths. Eosinophilic infiltration of the perineurial space has also been reported, particularly in those who have ingested alcohol recently, and the release of neurotoxins or pain mediators from eosinophils or mast cells has been suggested as a cause of pain.

Extrapancreatic causes of pain in chronic pancreatitis include stenosis or compression of the duodenum and common bile duct, and splenic vein thrombosis with splenic infarction. Duodenal and bile duct narrowing may result from fibrosis caused by long-standing pancreatic inflammation, from oedema in an acute exacerbation of pancreatitis, or from a pseudocyst. It is often difficult to determine the contribution of such extrapancreatic lesions to pain production but in general, pain is more often due to pancreatic causes. Duodenal ulceration is another potential cause of pain in chronic pancreatitis, the increased incidence of ulceration reflecting gastric hypersecretion and the diminished capacity of the pancreas to secrete bicarbonate and so neutralize acid in the duodenal bulb.

It is clear that there is no single cause of the pain of chronic pancreatitis and that there is great variation in its severity between individuals. Operations to relieve pain frequently have to be undertaken earlier in patients with alcohol-induced disease. Gross impairment of function with loss of secretory capacity and development of marked pancreatic calcification may be associated with increasing freedom from pain, but in many patients, an expectant approach is prevented by continuing severe pain and fears of opiate addiction.

Development of steatorrhoea indicates that pancreatic exocrine secretory capacity has fallen to less than 10 per cent of normal. The stools are bulky, pale, difficult to flush and have a particularly offensive and pervasive smell which causes considerable embarrassment. Inadvertent passage of oily droplets with staining of underwear can also be a problem. Watery diarrhoea is unusual because fat is not hydrolysed to fatty acids until it has been exposed to bacterial lipase in the colon; thus fatty acids are not present in the small intestine to trigger a secretory diarrhoea. If watery diarrhoea does occur in chronic pancreatitis it may reflect failure of pancreatic bicarbonate secretion with lowering of intraduodenal pH, precipitation of bile salts, and failure to solubilize fatty acids for absorption.

Excessive loss of dietary protein from failure to secrete proteases is seldom as marked as steatorrhoea, but contributes to the weight loss and muscle wasting seen in patients with severe chronic pancreatitis.

Failure of pancreatic endocrine function does not necessarily parallel the decline in exocrine function and occurs at a relatively late stage. When diabetes mellitus does develop it seldom responds to oral hypoglycaemic agents and exogenous insulin has to be prescribed. There is a real danger of insulin sensitivity as glucagon secretion is also reduced, with a resultant lowering of the capacity for gluconeogenesis. Insulin therapy must be monitored carefully and the patient made fully aware of the brittle nature of the diabetes. Diabetic ketoacidosis is rare in patients with chronic pancreatitis as there is usually enough residual insulin secretion to prevent release of fatty acids from adipose tissue and their subsequent metabolism in the liver to produce ketone bodies.

Nausea, vomiting, and dyspepsia are common. Anorexia may be present but is often a reflection of the fact that the patient is afraid to eat rather than true loss of appetite. Weight loss and malnutrition with vitamin deficiencies are particularly common in alcohol-induced chronic pancreatitis. Gastrointestinal, retroperitoneal, and intraperitoneal bleeding may arise from erosion of pancreatic and peripancreatic vessels involved in inflammation. Bleeding from peptic ulceration, gastritis, and duodenitis is not rare and oesophageal varices may be present as a consequence of splenic vein thrombosis and segmental portal hypertension.

Pleural effusion and ascites are more often complications of acute pancreatitis but both can occur in chronic pancreatitis.

While evidence of biliary obstruction is commonly present on liver biopsy in patients with chronic pancreatitis, hepatic cirrhosis affects less than 5 per cent of patients. There may be a link between inflammatory bowel disease and chronic pancreatitis but pancreatitis may simply result from inflammation at the papilla of Vater. Primary sclerosing cholangitis is linked to exocrine pancreatic insufficiency but its relationship to chronic pancreatitis per se is uncertain. An association between chronic pancreatitis and pancreatic cancer remains speculative but the two can coexist and pose diagnostic difficulty. The increased incidence of non-pancreatic cancer in patients with chronic pancreatitis almost certainly reflects the smoking habits and alcohol consumption of this patient population.

Advanced chronic pancreatitis is seldom difficult to diagnose in the presence of calcification, malabsorption, and diabetes. The diagnosis of early disease is more difficult, and suspicions raised by the history are seldom accompanied by findings on physical examination. Tests of pancreatic function are of limited value given the functional reserve of the pancreas, and more reliance is now placed on imaging investigations which can detect changes in pancreatic morphology.

An ideal test would detect pancreatic insufficiency before malabsorption became apparent, have a high specificity, and distinguish pancreatitis from pancreatic cancer. None of the tests available fulfils these criteria.

This is probably the best of the available tests of pancreatic function but it is now used rarely or not at all in many centres. Gastric contents are removed via a gastric tube and a duodenal tube is used to aspirate pancreatic secretions after stimulation by exogenous hormones. A marker such as ¹&sup4;C-polyethylene glycol may be used to ensure that at least 85 per cent of duodenal juice is recovered. Output of bicarbonate, trypsin, and amylase can be measured, but lipase determination is difficult technically. The test has a sensitivity of 75 to 90 per cent but attempts to improve its specificity beyond 80 to 90 per cent simply lower sensitivity.

This test used duodenal intubation to measure response to a test meal. It was even less sensitive than the secretin–cholecystokinin test and is now obsolete.

These remain in use although they are less sensitive and specific than the secretin–cholecystokinin test. In the bentiromide test, a synthetic tripeptide (N-benzoyl- l-tyrosyl- p-aminobenzoic acid) is given orally. In the presence of pancreatic chymotrypsin the tripeptide is cleaved to release p-aminobenzoic acid, which is absorbed from the small bowel, partially conjugated in the liver, and excreted in the urine. The pancreatolauryl test uses the same principle in that the ester fluorescein dilaurate is hydrolysed by pancreatic arylesterases to release fluorescein which is absorbed and excreted in the urine where it is detected spectrophotometrically. In patients with chronic pancreatitis judged to be severe on the basis of the secretin–cholecystokinin test, tubeless tests have a sensitivity of about 70 per cent; in mild to moderately severe disease this falls to below 50 per cent.

Many have been described but not fully evaluated. The secretion of pancreatic lipase can be assessed by blood radioactivity levels after ingestion of ¹&sup4;C-labelled triolein or by measuring breath ¹&sup4;CO&sub2;. Lactoferrin concentration (and lactoferrin: lipase ratio) in pancreatic juice is increased in chronic pancreatitis, and plasma pancreatic polypeptide levels have been correlated with enzyme secretion during fasting and after pancreatic stimulation. Serum trypsin-like immunoreactivity and pancreatic isoamylase concentrations during fasting are also lower in chronic pancreatitis. None of these tests is used in routine clinical practice.

Development of diabetes is assessed by testing for glycosuria, random blood glucose determinations, and an oral glucose tolerance test.

This is inexpensive, non-invasive and avoids ionizing radiation, and is the first investigation to be undertaken when chronic pancreatitis is suspected. Pancreatic size is assessed; in the early stages inflammation causes enlargement with an irregular outline, while in the advanced stages atrophy may ensue. There is often a heterogeneous pattern with increased echogenicity, and calculi charteristically produce very bright echoes. Pseudocysts and abscesses are readily detected. The pancreatic duct is frequently discernible and a diameter of more than 2 mm is regarded as abnormal. In normal individuals, injection of secretin produces an increase in duct diameter. In patients with chronic pancreatitis this may be prevented by surrounding fibrosis. Useful information may be obtained regarding primary disease in neighbouring organs, such as the liver and biliary system, which may be responsible for the patient's symptoms, and complications of pancreatitis affecting these organs, such as biliary obstruction and splenic infarction may be detected. Ultrasonography can also be used to guide percutaneous fine needle sampling for cytology or histology when the nature of a pancreatic mass lesion is uncertain, and can facilitate the aspiration or drainage of pseudocysts and abscesses. Percutaneous needle pancreatography can also be carried out under ultrasonographic control and Doppler ultrasonography can now be used to diagnose portal and splenic vein occlusion.

Since this investigation is more expensive than ultrasonography and requires exposure to ionizing radiation, it is not normally used as a first line investigation. It is, however, more sensitive than ultrasonography in the diagnosis of chronic pancreatitis (sensitivity 74 to 90 per cent as opposed to 60 to 70 per cent) and at least as specific (84 to 100 per cent versus 80 to 90 per cent respectively). Pancreatic size, density, and outline are readily assessed and CT scanning is the best method for detecting calculi. Pancreatic duct changes on CT scanning correlate well with those seen at endoscopic retrograde pancreatography, and duct size is best assessed after intravenous injection of contrast material. CT scanning also provides useful information about neighbouring organs and may distinguish between pancreatic cancer and chronic pancreatitis.

This is invasive, expensive, requires exposure to ionizing radiation and can give rise to complications which, on rare occasions, prove fatal. The Cambridge classification of radiological changes (Table 1) 389 is generally accepted and this investigation is widely regarded as the most accurate means of diagnosing chronic pancreatitis and distinguishing it from pancreatic cancer. Endoscopic retrograde pancreatography helps to define the extent and distribution of chronic pancreatitis and outlines the duct system. This information may prove invaluable in the selection of operative procedure if surgery is to be undertaken. Useful information may also be provided about involvement of the biliary tree. However, technical reasons may prevent cannulation of the pancreatic duct and ductal narrowing or obstruction may prevent full visualization of the duct system when contrast is injected. In such cases, ultrasonography and CT scanning may provide some of the missing information, but ultrasonographically guided percutaneous pancreatography or operative pancreatography may be needed to provide a comprehensive picture of duct morphology.

Some patients with symptoms and other evidence of chronic pancreatitis have little or no abnormality visible on endoscopic retrograde pancreatography. This has given rise to the controversial concept of ‘minimal change pancreatitis’.

Selective angiography and portal venography may be needed in selected patients, notably when portal or splenic vein thrombosis is suspected. Other imaging methods being evaluated at present include magnetic resonance imaging and endoscopic ultrasonography.

Assessment of nutritional status, full blood count, and liver function should be performed periodically, and it is worth carrying out random blood alcohol samples when continuing alcohol abuse is suspected.

The value of a good history cannot be over emphasized. When there is overt steatorrhoea, pancreatic function tests are unnecessary and one should proceed to ultrasonography and/or CT scanning. Even in the absence of steatorrhoea, pancreatic function tests (such as the secretin–cholecystokinin test or tubeless tests) are of dubious value. They are certainly useless in distinguishing between chronic pancreatitis and pancreatic cancer, and guided fine needle aspiration or biopsy should be undertaken whenever cancer is suspected. The presence of pancreatic calcification does not exclude cancer and pancreatitis and cancer may coexist. Endoscopic retrograde cholangiopancreatography should be considered once the pancreas has been imaged by ultrasonography and/or CT scanning, particularly when the diagnosis remains in doubt, duct morphology has not been defined, and surgery is contemplated.

The diagnosis of chronic pancreatitis is not in itself an indication for surgery. Some patients have little or no pain, while in others pain can be controlled without recourse to opiates. Surgery does not restore exocrine and endocrine function; indeed pancreatic resection may cause marked deterioration and precipitate the onset of diabetes. Chronic pancreatitis is often difficult to manage and when there are problems arising from alcoholism, social deprivation, and personality disorders it may be impossible to disentangle cause from effect. Good rapport and mutual trust are essential for successful long-term management and require frequent outpatient consultation with periods of inpatient assessment if appropriate.

Effective pain relief is usually the most difficult aspect of management. Patients vary considerably in pain threshold and pain perception and many are already taking opiates when first referred for surgical consultation. They should be encouraged to abstain from alcohol, take frequent small meals rather than infrequent large ones, and avoid overtly fatty foods. Liquid meals and elemental diets are no longer prescribed in attempts to reduce pancreatic stimulation: liquids augment the gastric phase of pancreatic secretion while elemental diets stimulate cholecystokinin release. Pancreatic extracts are used to treat steatorrhoea and probably reduce pain by providing intraluminal trypsin which blocks release of cholecystokinin from the duodenal mucosa. Histamine H&sub2;-receptor antagonists such as cimetidine or ranitidine may prevent or eradicate duodenitis and peptic ulceration, and theoretically reduce pancreatic stimulation by diminishing duodenal release of secretin.

Analgesics must be prescribed carefully, given the long-term nature of the problem and risk of opiate addiction. Unfortunately, non-opioid drugs such as aspirin and paracetamol are more effective at relieving musculoskeletal pain than visceral pain, and opiates are frequently essential. All opioid drugs may cause drowsiness, nausea and vomiting, and constipation (although this may be regarded as beneficial in patients with steatorrhoea). Sublingual bruprenorphine (200–400 &mgr;g every 6–8 h), and oral dihydrocodeine tartrate (30 mg every 4–6 h) or pentazocine (25–100 mg every 3–4 h after food) may give satisfactory pain control without recourse to injection. More severe exacerbations of pain may require injection of morphine (10–30 mg) or pethidine (50–100 mg) and long-lasting severe pain may need oral pethidine (50–150 mg every 4 h) or long-acting morphine preparations (such as morphine sulphate, MST Continus® 10–60 mg twice daily). Whenever such powerful opioid drugs are having to be prescribed frequently and/or continuously, serious consideration must be given to surgery.

The sensory sympathetic nerves which transmit pain impulses from the pancreas synapse in the coeliac ganglion on their way to the splanchnic nerves. Attempts to relieve the pain of chronic pancreatitis by CT guided injection of alcohol in and around the coeliac plexus have had extremely variable results and at best give pain relief for a few months. Coeliac blockade is no longer used in most centres.

In theory, provision of exogenous lipase supplements with every meal should readily eliminate steatorrhoea. In practice steatorrhoea may be difficult to abolish as gastric acid destroys up to 90 per cent of such enzyme activity by the time food reaches the distal duodenum. Antacids or H&sub2;-receptor antagonists were once used to combat this problem. However, modern pancreatic exocrine supplements (e.g. Creon; one capsule of which contains 8000 units lipase, 9000 units amylase, and 210 units protease) consist of enteric coated granules of pancreatic extract contained in an outer gelatin capsule. The gelatin dissolves with a few minutes in the acidic stomach whereas the granules break down in the duodenum when luminal pH rises above pH 5.5. Microsphere preparations are more expensive than simpler preparations of pancreatic extracts but are well tolerated. The initial dose is 1 to 2 capsules with meals, rising to 5 to 15 capsules each day. If steatorrhoea persists, H&sub2;-receptor antagonists should be prescribed. If steatorrhoea is still troublesome, alternative explanations (such as bacterial overgrowth in the small intestine or bile salt insufficiency) should be considered. If all else fails, fat intake should be reduced to below 40 g per day and the diet supplemented with medium chain triglyceride oil which requires less lipase for cleavage.

Insulin is needed if diabetes mellitus develops. As indicated earlier, the risk of hypoglycaemia is much greater than that of ketoacidosis.

Nutritional status requires careful supervision. An adequate intake of calories and vitamins is essential, particular attention being paid to intake in alcoholic patients. Recent suggestions that antioxidant supplements may be of value in pain relief require further evaluation.

Pain is the most common indication for surgery. The decision to operate is only taken after full trial of medical therapy and comprehensive evaluation of the patient and his domestic situation. The patient must not be ‘talked into’ undergoing surgery and should be aware that many patients do not require surgery and that pain may gradually abate with time. He should have no illusions about the prospects following operation. Exocrine and endocrine function do not return to normal and arrest or slowing of functional deterioration is a realistic goal if a drainage operation is proposed. If resection is required, brittle lifelong diabetes may be precipitated and is inevitable after total pancreatectomy. Surgery does not always cure or improve pain and, even after total pancreatectomy, up to 20 per cent of patients still experience pain. A realistic guideline is that some 70 per cent of patients are still pain free or improved 5 years following operation. Many surgeons refuse to operate on alcoholics who are still drinking, and all agree that the prognosis after operation is much better in those who abstain.

Other factors which may indicate the need for surgery are the development of complications (see below) and uncertainty regarding the presence of pancreatic cancer.

This is the standard operation when the pancreatic duct system is dilated to more than 7 to 8 mm in diameter. A complete pancreatogram is advisable and operative pancreatography (using needle puncture of the duct in the body or tail of the pancreas or transduodenal cannulation) may be needed if endoscopic retrograde pancreatography has been unsuccessful. The pancreas is exposed fully by division of the gastrocolic ligament and the pancreatic duct is opened fully along its length (Fig. 1) 1329. A minimum length of pancreatico-jejunal anastomosis of 10 cm is advised for long-term patency; the longer the anastomosis the better. Calculi are removed from the duct system and a Roux loop of proximal jejunum is brought through a window in the transverse mesocolon and anastomosed to the pancreatic duct in side-to-side manner. It is not necessary to remove the spleen during this operation. The Roux loop is orientated so that its end lies on the tail of the pancreas; this allows the same loop to be used for anastomosis to the biliary system should biliary obstruction develop. If biliary obstruction is already present, a side-to-side hepaticojejunostomy is undertaken at the time of pancreaticojejunostomy. Pseudocysts can also be drained into the Roux loop.

Alternatives to longitudinal pancreaticojejunostomy include caudal pancreaticojejunostomy and pancreaticogastrostomy. The former operation involves anastomosing a loop of jejunum to the cut surface of the pancreas after removing the tail; it is now obsolete. Pancreaticogastrostomy still has adherents but there is no objective evidence for its superiority over the conventional operation.

This may be performed with or without removal of calculi from the pancreatic duct, and is of debatable value. Benefit has been claimed in small series of selected patients but for the vast majority it seems inherently unlikely that pancreatitis arises from obstruction of the terminal portion of the duct and that sphincteroplasty will give long-term success. This said, there is now a growing number of reports of treatment of chronic pancreatitis by endoscopic papillotomy or balloon dilatation, accompanied in some cases by insertion of a stent and destruction of calculi by extracorporeal shock wave lithotripsy. The long-term results are yet to be defined but the approach may defer or avoid operation in selected patients.

Endoscopic and surgical means can be used in this manner to treat chronic pancreatitis associated with pancreas divisum. In the main, the results have been disappointing, calling into question the significance of pancreas divisum as an aetiological factor.

Removal of 40 to 95 per cent of the pancreas was at one time popular, but its popularity waned with the realization that many patients still had pain after surgery and that marked exocrine and endocrine insufficiency often resulted. There is still a place for distal resection, conserving the spleen if possible, in patients in whom chronic inflammation appears to be confined to the distal pancreas.

This is regarded by many as the standard operation when the duct system is not sufficiently dilated for pancreaticojejunostomy. The head of the gland is often the most inflamed part, the operation conserves useful exocrine and endocrine function by retaining the body and tail, any biliary tract obstruction is dealt with, and the risk of leaving occult cancer in the head of the pancreas is avoided. The gallbladder should be removed to avoid subsequent gallstone formation and the jejunum is anastomosed to the common hepatic duct rather than common bile duct to reduce the risk of ischaemic breakdown of the anastomosis. Opinions vary as to whether antrectomy should be performed or whether the first part of the duodenum should be transected with preservation of the pylorus and entire stomach. One variant preserves the stomach and entire duodenum. The case for pylorus-preserving and duodenum-preserving pancreaticoduodenectomy rests on retention of more normal gastrointestinal physiology and function, but has yet to be accepted universally. Truncal vagotomy was once an integral part of the Whipple operation but the risk of marginal ulceration is now so small that few surgeons divide the vagi. Most surgeons attempt to retain exocrine function by anastomosing the cut surface of the remaining pancreas to the jejunum. Others attempt mucosa-to-mucosa apposition between the pancreatic duct and jejunum while some doubt the value of attempting to retain exocrine function and staple the transected pancreas or occlude its duct system.

This is a last resort given the inevitable diabetes and steatorrhoea which result. Most patients coming to total pancreatectomy have already failed to benefit from lesser surgical procedures and it must be re-emphasized that some 20 per cent of patients still complain of pain following total extirpation. Pylorus-preserving and duodenum-preserving forms of total pancreatectomy have been used. One advantage sometimes claimed for total pancreatectomy is that it avoids the need for anastomosis between residual pancreas and jejunum. However, this anastomosis is not as dangerous in chronic pancreatitis as it is in cancer surgery. In chronic pancreatitis the gland is fibrous and firm rather than friable, and the reduced exocrine secretory capacity lowers the risk of leakage and fistula formation.

Attempts to relieve pain by excision of the coeliac ganglion and other forms of neurectomy have generally proved disappointing. Left transthoracic splanchnicectomy and bilateral truncal vagotomy has recently been advocated but requires further evaluation.

The ideal operation should have no operative mortality, provide permanent relief of pain, and conserve pancreatic exocrine and endocrine function. Additional aims would be minimal interference with normal digestive function and elimination of the risk of failing to remove occult pancreatic cancer. All operations should now carry a risk of operative mortality near to zero if performed in specialist hands. Pancreaticojejunostomy is safer than resection and offers a good prospect of pain relief with maximal conservation of pancreatic function. Unfortunately only some 20 to 30 per cent of patients have a duct system which is sufficiently dilated for this operation. It remains to be seen whether endoscopic procedures will retain a place in management as an alternative to surgical drainage.

Resection has become safer with improved operative and perioperative care but still carries a greater risk than drainage surgery. Tissue planes are frequently destroyed by inflammation, and operative damage to major vessels (particularly the portal and superior mesenteric vein) poses a real hazard. The greater the resection, the greater the risk of exocrine and endocrine insufficiency. Distal pancreatectomy is indicated when disease is confined to or is maximal in the body and tail and can be combined with drainage of the duct system of the remaining pancreas. The Whipple operation (or its variants) is now used in the majority of patients, while total pancreatectomy is reserved for patients in whom lesser procedures fail to relieve pain.

The availability of ultrasonography and CT scanning has improved our understanding of the natural history of pseudocysts in chronic pancreatitis. As a general rule pseudocysts which exceed 5 cm in diameter and which have been present for more than 6 weeks are unlikely to resolve spontaneously. When treatment is needed, internal drainage is safer and more effective than external drainage or resection. Biopsy of the pseudocyst wall is always advisable at surgery as cystic pancreatic neoplasms can cause confusion in diagnosis. A Roux loop of jejunum is usually recommended for internal drainage although on occasions the stomach or duodenum may be used. If a fluid collection proves to be an abscess, external drainage is advisable. False aneurysms can masquerade as pseudocysts and erosion of major vessels may cause bleeding into a pre-existing pseudocyst. It is always advisable to have a high index of suspicion, sample the cyst contents by needling before incision, and consider preoperative angiography when any doubt exists.

Transient jaundice and elevation of serum alkaline phosphatase levels is common in chronic pancreatitis. Ultrasonography and, if necessary, endoscopic retrograde cholangiopancreatography can be used to define biliary morphology, and liver biopsy is indicated if there is long-standing derangement of liver function. However, alcoholic hepatitis and cirrhosis are surprisingly uncommon, even in patients with chronic alcohol-associated pancreatitis, secondary biliary cirrhosis is rare, and most patients merely have histopathological evidence of biliary obstruction with or without cholangitis. When the bile duct is narrowed a long segment of stenosis in the retropancreatic bile duct is usually found although hour-glass constriction or lateral displacement may occur.

Bile duct obstruction in chronic pancreatitis is usually due to fibrosis of the surrounding pancreas rather than acute inflammation or pressure from a pseudocyst. Given the long-term risk of cholangitis and secondary biliary stenosis, fibrotic narrowing of the bile duct requires surgical treatment. Such surgery to relieve biliary obstruction is needed in 10 to 20 per cent of patients with chronic pancreatitis. Resection of the head of the pancreas may also be indicated because of pain, but if resection is not needed biliary drainage can be achieved by hepaticojejunostomy or choledochoduodenostomy. Cholecystojejunostomy is not recommended as it fails to ensure long-term bile flow and leaves the patients at risk of developing gallbladder disease. Temporary drainage can be obtained by endoscopic biliary stenting if there is a need to avoid operation.

Transient duodenal obstruction may complicate an exacerbation of inflammation in the head of the pancreas. Chronic fibrotic obstruction is a relatively rare complication of chronic pancreatitis (<2 per cent of cases) and endoscopy is essential to exclude peptic ulceration, Crohn's disease, and pancreatic or periampullary cancer. Persisting fibrotic obstruction demands surgical relief. Gastroenterostomy is used if duodenal obstruction is the sole indication for surgery or when pancreaticojejunostomy is also needed. Where there is painful disease in the head of the pancreas and no duct dilatation, the Whipple operation is used.

Colonic obstruction is rare; when it occurs it is usually transient, involving the distal colon or splenic flexure. Operation is only indicated if obstruction persists for more than a few weeks.

Arterial haemorrhage may result from necrosis of the vessel wall in areas of inflammation and pseudocyst formation. The splenic, gastroduodenal, pancreaticoduodenal, pancreatic, gastric, and hepatic arteries may be involved, in order of frequency. Rupture can result in pseudoaneurysm formation and bleeding can occur into the pancreatic duct, pseudocyst, retroperitoneal tissues, or peritoneal cavity. Endoscopy and selective angiography are useful investigations and although surgery is usually indicated, therapeutic embolization may be life saving in high risk patients.

Thrombosis of the splenic vein, and less frequently of the portal and superior mesenteric vein, can be asymptomatic. Alternatively, the patient may bleed from gastric and oesophageal varices or present with a large spleen, leucocytopenia, and thrombocytopenia. Haemorrhage from colonic varices is exceptional. When variceal bleeding occurs after splenic vein occlusion, splenectomy is the treatment of choice. A preoperative arteriogram and venogram is advisable, partly to define the problem and partly to embolize the splenic artery to reduce collateral bleeding at operation.

Leakage of pancreatic juice from rupture of a pseudocyst or inflammation and necrosis of the pancreatic duct may result in rapid development of ascites and pleural effusions. The diagnosis is confirmed if the fluid protein content exceeds 25 g/l and if the amylase concentration of the fluid exceeds that of the serum. Conservative treatment with parenteral nutrition and the somatostatin analogue SMS201–905 (to reduce pancreatic secretion) may be employed initially but is seldom successful. Surgery should not be postponed beyond 2 to 3 weeks. Preliminary endoscopic retrograde pancreatography is invaluable to define the site of leakage. Leaks in the tail are usually treated by distal pancreatectomy, those elsewhere by pancreaticojejunostomy.

Internal pancreatic fistulae involving the gut may not require surgical intervention and are often discovered as incidental findings at endoscopic retrograde pancreatography. Pancreaticocutaneous fistulae frequently follow external drainage of pseudocysts or may complicate pancreatic surgery. They often respond to conservative measures but surgical resection of the involved pancreas may be needed.

Most workers accept that hereditary chronic pancreatitis predisposes to pancreatic cancer and there is now some epidemiological evidence suggesting that non-hereditary chronic pancreatitis may also be a premalignant condition. Definition of the relationship between chronic inflammation and cancer is difficult as the two conditions frequently coexist and an occluding cancer may give rise to inflammation.

The prognosis in chronic pancreatitis depends on the frequency and severity of the attacks, need for surgery, and the development of complications, notably diabetes. Alcoholics who fail to stop drinking undoubtedly fare worse than non-alcoholics and alcoholics who abstain. Given the importance of alcohol as an aetiological factor and its associated personality and social upset, it is hardly surprising that patients with chronic pancreatitis have a lower life expectancy than the general population. The complications of the disease, drug addiction, depression, brittle diabetes, malnutrition, and increased risk of non-pancreatic cancer combine to reduce longevity. Reported survival rates vary greatly according to the nature of the series of patients studied. The cumulative survival rate based on life table analysis in alcoholics with or without surgery is about 50 per cent at 20 to 24 years from onset of the disease; non-alcoholic patients have survival rates which are some 20 per cent higher.

Ammann RW, Akovbiantz A, Largiader F, Schueler G. Course and outcome of chronic pancreatitis. Longitudinal study of a mixed medical-surgical series of 245 patients. Gastroenterology 1984; 86: 820–8.

Axon AT, Classen M, Cotton PB, Cremer M, Freeny PC, Lees WR. Pancreatography in chronic pancreatitis: international definitions. Gut 1984; 25: 1107–12.

Banks PA. Medical strategy in chronic pancreatitis. In: Carter DC, Warshaw AL, eds. Pancreatitis. Edinburgh: Churchill Livingstone, 1990; 133–47.

Bockman DE, Buchler M, Malfertheiner P, Beger HG. Analysis of nerves in chronic pancreatitis. Gastroenterology 1988; 94: 1459–69.

Braasch JW, Vito L, Nugent W. Total pancreatectomy for end-stage chronic pancreatitis. Ann Surg 1978; 188: 317–22.

Cooper MJ, et al. Total pancreatectomy for chronic pancreatitis. Br J Surg 1987; 74: 912–15.

Copenhagen Pancreatitis Study. An interim report from a prospective epidemiological multicenter study. Scand J Gastroenterol 1981; 116: 305–12.

Eckhauser FE, Strodel WE, Knol JA, Harper M, Turcotte JG. Near-total pancreatectomy for chronic pancreatitis. Surgery 1984; 96: 599–607.

Grace PA, Pitt HA, Longmire WP. Pylorus preserving pancreatoduodenectomy: an overview. Br J Surg 1990; 77: 968–74.

Greenlee HB, Prinz RA, Aranha GV. Long-term results of side-to-side pancreatico-jejunostomy. World J Surg 1990; 14: 80–4.

Lambert MA, Linehan IP, Russell RCG. Duodenum preserving total pancreatectomy for end-stage chronic pancreatitis. Br J Surg 1987; 74: 35–9.

Niederau C, Grendell JH. Diagnosis of chronic pancreatitis. Gastroenterology 1985; 88: 1973–95.

Sarles H, Bernard JP, Gullo L. Pathogenesis of chronic pancreatitis. Gut 1990; 31: 629–32.

Sauerbruch T, Holl J, Sackmann M, Paumgartner G. Extracorporeal shock wave lithotripsy of pancreatic stones. Gut 1989; 30: 1406–11.

Singh SM, Reber HA. The pathology of chronic pancreatitis. World J Surg 1990; 14: 2–10.

Stone HH, Chauvin EJ. Pancreatic denervation for pain relief in chronic alcohol associated pancreatitis. Br J Surg 1990; 77: 303–5.

Wilson C, et al. Hepatobiliary complications in chronic pancreatitis. Gut 1989; 30: 520–7.

Worning H. Incidence and prevalence of chronic pancreatitis. In: Beger H, Buchler M, Ditschuneit H, Malfertheiner P, eds. Chronic Pancreatitis. Berlin: Springer-Verlag, 1990: 8–14.