{49234} chronic pancreatitis; pale white is scar
{08853} chronic pancreatitis; extensive scarring
{20279} chronic pancreatitis in an alcoholic, nice protein plug
{46283} chronic pancreatitis with calculi
{49230} burned out alcoholic chronic pancreatitis, with calculi; the tube along the bottom is the splenic artery, twisting in and out

In the late stages, patients can expect to have malabsorption (steatorrhea, weight loss), and perhaps a pseudocyst. "Big Robbins" does not mention the worst problem that many of these people have -- chronic severe pain from involvement of the sensory nerves around the celiac plexus.

As you must have noticed, there is little justification for calling this "chronic pancreatitis", except that it may be chronically painful. This misnomer was canonized in the Marseilles-Rome criteria of 1988 (Scand. J. Gastroent. 24: 641, 1989).

Chronic obstructive pancreatitis, a slightly different entity from the above, follows obstruction of the pancreatic duct (gallstone, surgeon's mishap). In this situation, there is selective atrophy of acini around the head of the pancreas. Surgeons can repair the ampulla if that is the problem.

* Exactly how the cells die remains obscure; there are conflicting animal models (apoptosis vs. inflammation/necrosis): Gastroenterology 110: 875, 1996.

* The most interesting work in chronic pancreatitis recently is the discovery that antioxidants seem to help a lot: Am. J. Gastro. 91: 1558, 1996, others.

Future pathologists: Here's how to distinguish these two entities (after World. J. Surg. 14: 2, 1990):

NON-MALIGNANT MASSES

* Cysts of the pancreas are rare. They are seen (in a minority of cases) in two probable anti-oncogene-deletion syndromes (Von Hippel-Lindau, adult polycystic kidney disease).

Pancreatic pseudocysts (see above) are common after acute pancreatitis from alcoholism or trauma, or from any other cause.

Benign tumors of the pancreas are uncommon. They are almost always some variant of adenoma. Cystadenomas are the most common. Leave it to the pathologist to decide whether a particular tumor has malignant potential Cancer 71: 82, 1993.

* Intraductal papillary neoplasms: Cancer 72: 689, 1993 (can express both exocrine and endocrine markers, which is not surprising, since both come from ducts); Ann. Surg. 223: 141, 1996. Dysplastic ones can probably turn into cancer.

{49238} cystadenoma of the pancreas; has been cut in half and opened; spleen at right

CANCER OF THE PANCREAS ("cank of the pank", "the dismal disease", etc.; Lancet 363:1049, 2004; Disease-A-Month 50: 545, 2004)

This dread cancer accounts for about 5% of U.S. cancer deaths; the incidence has tripled in the past 50 years "because of smoking and chemicals" (I wonder).

* Famous victims in recent memory include Billy Carter, Jack Benny, Dizzy Gillespie, Rex Harrison, and Michael Landon.

Virtually all cancers of the pancreas are adenocarcinomas arising from the ducts.

Like adenocarcinomas anywhere else, you can spot them because they make little glands and/or are secretory- product (i.e., mucin)-positive.

* Truly hardcore future pathologists will want to read about MUC4 as a stainable marker for pancreatic neoplasia, especially as it grows nastier: Am. J. Clin. Path. 117: 791, 2002.

{08851} adenocarcinoma of pancreas; no normal pancreas on the slide; some glands are more anaplastic than others;
{26003} adenocarcinoma of pancreas; mucin-producer (pale apical cytoplasm, sharp borders)

Cancer of the pancreas WebPath Case of the Week

Cancer of the pancreas Primary and liver metastases KU Collection

Future pathologists and surgeons: Cancer of the pancreas and chronic pancreatitis are hard to tell apart. A certain percentage of false-positive diagnoses of cancer of the pancreas, and a certain number of Whipple procedures for those without cancer of the pancreas, is acceptable: Br. J. Surg. 81: 585, 1994.

Future pathologists only: The in-situ evolution of pancreatic cancer has been well-studied. See Arch. Path. Lab. Med. 118: 227, 1994.

* For some reason, it is not uncommon to see real osteoclasts and even osseous metaplasia here (J. Clin. Path. 47: 372, 1994; Arch. Path. Lab. Med. 120: 306, 1996); these are still carcinomas.

Risk factors include (1) smoking (3x the normal risk), (2) exposure to chemicals (the disease is supposedly more common among both chemists and garage workers), (3) hereditary pancreatitis (huge risk), and (4) Some of the anti-oncogene deletion syndromes.

"Big Robbins" links it to the notable carcinogens naphthylamine and benzidine, and the hoopla over nitrosamines in food was related to their link to cancer of the pancreas in experimental animals.

* Questionable risk factors include alcohol consumption, high-fat diet ("cholecystokinin must be a promoter"; tough to believe if you think pancreatic acinar cells don't ordinarily divide), coffee drinking, obesity, and pernicious anemia. All of these are now pretty much discredited (see for example Cancer 67: 2664, 1991; the coffee crock discredited Cancer Epidem. 10: 429, 2001, several others uniformly negative).

Anti-oncogene deletion syndromes placing people at excess risk for pancreatic cancer include BRCA2 (breast-and-ovary), the Lynch hereditary nonpolyposis coli cancer syndromes (hMSH2, hMLH1), Peutz-Jegher's (STK1/LKB1; Big Robbin's claim of a x130 increased risk can't be right), and the p16/CDKN2A pancreatic cancer/melanoma syndrome (NEJM 350: 2623, 2004).

Hereditary pancreatitis in particular gives at least a 40% risk of getting pancreatic cancer (Med. Clin. N.A. 84: 719, 2000).

* Concerns about exposure to particular pesticides keep coming up, but the whole business remains very soft -- the ones that "mutate k-ras in lab animals" aren't the ones that "are associated with a 5x increased risk of cancer of the pancreas in workers", etc., etc. To become confused, see Env. Health Perspect. 111: 724, 2003; Lancet 354: 2125, 1999.

Whatever the environmental "cause", most (or maybe all) cancers of the exocrine pancreas have mutated k-ras at hot-spot codon 12. This can be detected on fine-needle aspirate material, and by PCR in pancreatic fluid (Cancer 73: 1589, 1994; Am. J. Path. 144: 889, 1994) or stool (ooh, Cancer Res. 54: 3568, 1994) or smears (Am. J. Clin. Path. 105: 257 & 321, 1996). Smoking seems to cause this mutation (Cancer 85: 326, 1999).

The distribution of cancers in "Big Robbins" is reasonable:

60% head
15% body
5% tail
20% too late to tell

Patients come in with back pain (why?), jaundice, weight loss, GI upsets, depression (very typical, and poorly understood), and/or migratory thrombophlebitis ("Trousseau's other sign"; the mechanism of the distinctive paraneoplastic problem is unknown).

The size of the cancer depends on the stage at which it is detected. Those in the head may be found early because they produce jaundice. Those in the body and tail will be detected late.

There's a serum tumor marker, CA-19-9 (Gut 35: 707, 1994, many others). * Another, for cyst fluid: CA-72-4 (Ann. Surg. 219: 131, 1994).

Future surgeons: Courvoisier's law states that a distended gall bladder in a patient with obstructive jaundice means cancer (pancreas, common bile duct). Obstruction due to a gallstone in the common bile duct will not result in a distended gallbladder, because the gallbladder would be heavily scarred-up from years of cholelithiasis. This works most of the time, though you would never rely on it.

Gung-ho surgeons may try to resect a tumor in the head of the pancreas ("Whipple procedure"; * for the pylorus-sparing technique see J. Am. Col. Surg. 178: 443, 1994; for the Hopkins study, which hails 11 cures out of 201 surgeries as an enormous improvement, see Ann. Surg. 221: 721, 1995. Desperate diseases require desperate remedies).

The overall 5-year survival rate in cancer of the pancreas is about 1%; typical survival is about six months (Gut 31: 494, 1990; J. Am. Col. Surg. 179: 38, 1994).

These patients often have diabetes, and the cause is insulin resistance. This now appears to be due to massive production of amylin (IAPP; NEJM 330: 313, 1994; Gastroenterology 114: 130, 1998); probably the amylin is produced by the islets in response to one or more factors produced by the tumor itself (J. Clin. Endo. Metab. 85: 1232, 2000). This probably explains the well-known "link" between pancreatic cancer and diabetes, and it now appears that only new-onset (i.e., less than three years) diabetes is a "risk factor" (NEJM 331: 81, 1994).

* A recent British euthanasia case involved intractable pain from cancer of the pancreas: Lancet 335: 719, 1990 ("not guilty"; in my opinion this is a triumph of humanity and common sense; you may disagree).

Adenocarcinoma of the pancreas typically metastasizes to lymphatics, and blood-borne metastases to the liver are typically massive.

* Acinar carcinoma of the pancreas is a rare PAS-positive tumor, usually of younger people. It presents stippled cells that stain for amylase, lipase, and chymotrypsin, and often elaborates lipase into the blood (which may produce subcutaneous fat necrosis!) Molecular genetics: Am. J. Path. 160: 953, 2002.

CLASSIFYING DIABETES MELLITUS AND RELATED CONDITIONS

Diabetes Mellitus Text and pictures From "Big Robbins"

Pathology of Diabetes WebPath Tutorial

"Diabetes" literally means "siphon", because of the osmotic diuresis produced by the glycosuria. This was known all-too-well to Hippocrates, who may have named it.

Diabetes mellitus (MELL-uh-tuss, please) is "a chronic disorder of carbohydrate, fat, and protein metabolism characterized in its fully expressed clinical form by an absolute or relative insulin deficiency, fasting hyperglycemia, glycosuria, and a striking tendency toward the development of atherosclerosis, microangiopathy, nephropathy, and neuropathy" (old Big Robbins).

Diabetes is our commonest serious metabolic disease, affecting maybe 5% of the population. On the average, it takes 15 years off the patient's life (JAMA 285: 628, 2001) and accounts for a tremendous amount health care expenses.

You will need to know the terminology (which is often not used correctly):

Diabetes mellitus ("overt diabetes", "manifest diabetes", etc.): the patient has...

• signs and/or symptoms of diabetes plus any serum glucose of 200 mg/dL of more, or...
• a serum glucose of 200 mg/dL or more at the 2-hour point of a glucose tolerance test, or...
• elevated fasting blood sugar over 126 mg/dL on two occasions with the patient in his or her usual state of health.

The fasting criteria are down from 140 mg/dL (Am. Fam. Phys. 58: 1355, 1998). It identifies people at risk for health problems from hyperglycemia.

There's talk today about distinguishing "impaired fasting glucose" (IFG, i.e., 110-125 mg/dL) and "impaired glucose tolerance" (IGT, i.e., 121-179 mg/dL at the two-hour mark). See Arch. Int. Med. 161: 397, 2001.

Type I diabetes and Type II diabetes (below) are sometimes called "primary diabetes", since they seem to be genetic diseases in their own right.

Secondary diabetes is said to exist when the metabolic disturbances are the result of some other identifiable illness, injury, molecular abnormality, etc., etc.

Impaired glucose tolerance ("glucose intolerance", "subclinical diabetes", "asymptomatic diabetes", "chemical diabetes", "latent diabetes"): fasting blood sugar is normal, but a glucose tolerance test is abnormal. Current recommendations are NOT go looking for this: Am. J. Med. 105(1A): 15S, 1998.

Gestational diabetes mellitus: diabetes mellitus first appearing during pregnancy, and perhaps disappearing when the pregnancy ends.

"Previous Abnormality of Glucose Tolerance" ("prediabetes", "latent diabetes"): the patient once had measurable glucose intolerance (as, when she was pregnant), but is chemically normal now (but may be at risk for future diabetes mellitus, depending on the circumstances).

"Potential Abnormality of Glucose Tolerance" ("prediabetes"): the monozygotic twin of a type II diabetic, or (less justifiably) someone else with a strong family history.

Not diabetes: Glucose intolerance only under some obvious physiologic stress (myocardial infarction, pneumonia, severe burns, terror of venipuncture, etc.) Mostly an epinephrine effect; probably cortisol contributes as well.

The new system recommended by the WHO and the American Diabetes Association (Br. Med. J. 317: 359, 1998):
Type I: autoimmune and idiopathic situations in which beta cell function is essentially gone, and there is an absolute insulin lack;

Type II: defects in insulin secretion, and/or a relative lack of insulin, and/or insulin resistance;

Type III: damage to the whole pancreas (old pancreatitis, cystic fibrosis -- diabetes from CF is uncommon but happens J. Ped. 142: 97, 2003) and autosomal dominant genetic syndromes (I'm clueless as to why these are lumped together);

Type IV: Gestational diabetes

Insulitis
Type I diabetes
WebPath photo

PRIMARY DIABETES TYPE I ("juvenile onset", "labile", "ketoacidosis-prone", "insulin-dependent"): 10% of diabetics.

One person in 300 in the U.S. gets this kind of diabetes (rates vary considerably from nation to nation; * rates are higher at higher latitudes).

Typical case: A child (average age twelve years, but we now know you can get the disease at any age) presents with polyuria, polydipsia, and polyphagia of relatively sudden onset. The child is found to have very high blood glucose levels causing osmotic diuresis.

Before the era of injectable insulin, diabetic ketoacidosis (DKA) and death followed in short order.

You remember the pathophysiology of ketoacidosis from your physiology course. Future clinicians: Ketoacids impart the familiar "rotten apples" sweetness to these patients' breath.

Today, the child looks forward to a period of fairly good health while taking injectable insulin, checking blood glucose several times a day with chemical strips and a reflectance meter.

After 10-15 years, unless control is good, the diabetic starts to suffer with infections, eye problems, peripheral neuropathy, gangrene of the lower extremities, kidney disease, stroke, and coronary atherosclerosis.

Historically, death usually came about forty years after onset as the result of a myocardial infarction. By this time, 50% of patients had lost their kidneys, and nearly as many were blind, stroked out, legless, and/or in chronic pain from neuropathy. A well-treated, compliant diabetic typically does better today.

The essential lesion in type I diabetes is a severe absolute lack of insulin.

* Only half of patients have any evidence of insulin production (measure C-peptide in serum).

Insulin deficiency and hyperglycemia explain the presentation but do not explain the later complications of the disease.

"Type I diabetes is a genetically programmed, chronic autoimmune disease" (NEJM 314: 1360, 1986, an early review; update Nature 351: 519, 1991), with the acute-symptomatic phase sometimes triggered by an acute viral illness.

In other words, the etiology is kind-of-complex.

Genetic factors:

Siblings of those with Type I diabetes are at increased risk (25x).

Identical twins of those with Type I diabetes have a 50% chance of eventually getting it also.

Type I diabetes is strongly associated with HLA-related antigens DR3 and DR4. (* If one has the misfortune to have both, it's even worse.... The former association with some HLA-B antigens was due to their linkage to DR3 and DR4; and currently, it appears that the also-linked DQ is the closest important site.)

* As is so common when the immune system attacks gland parenchyma, the beta cells of these patients express HLA class II histocompatibility antigens. No one knows whether this is cause or effect.

* The molecular defect that permits type I diabetes to occur seems to be homozygous absence of aspartic acid in position 57 of the HLA class II DQ chain (Nature 329: 599, 1987; Nature 333: 710, 1988), at least in the U.S. Update on HLA links: J. Clin. Endo. Metab. 89: 4037, 2004; they vary tremendously from nation to nation: J. Clin. Endo. Metab. 90: 5104, 2005. The famous locus IDDM1, where certain polymorphisms give a risk for type I diabetes, is a component of the HLA system (Diabetes 50: 1200, 2001).

The gene IDDM2 ("implicated in diabetes melitus") is a complicated, highly variable tandem repeat adjacent to the real insulin gene. Two variants are strongly linked to type I diabetes (update Diabetes 53: 1884, 2004).

The animal model of autoimmune diabetes is the non-obese diabetic mouse, which gets that way because of genes at three (or more) loci (Nature 353: 260, 1991; J. Imm. 152: 204, 1994). Update J. Immuno. 169: 6617, 2002; to date; the exact reasons for the famous mouse's diabetes remain elusive.

* The BB (formerly BB/W) rat is a strain discovered in 1977. These rats have autoimmune insulitis, and the majority develop acute-onset type I diabetes. They helped us find the IDDM1 and IDDM2 loci (Acta. Diabet. 35: 109, 1998).

Autoimmune factors:

Several types of IgG anti-beta-cell antibodies occur. One or more is present in the vast majority of type I diabetics the acute phase (contrast 0.5% in healthy people). It is now quite clear that they are etiologic, and that they are usually present before age 2 in children destined to get type I diabetes (Ann. Int. Med. 140: 882, 2004).

Their specificities include anti-glutamic acid decarboxylase (Nature 347: 151, 1990; NEJM 322: 1555, 1990; Lancet 341: 1378 & 1383, 1993; diabetogenic epitope Lancet 343: 1607, 1994; true both of NOD mice and people: Nature 366: 69 & 72, 1993).

There is also cell-mediated immunity directed against beta cells in most patients who have been studied. Again, the autoantigen is glutamic acid decarboxylase. Update Nature 391: 177, 1998.

There was generally a dense lymphocytic infiltrate in the islets of patients dying in the acute phase (rare nowadays).

* Maybe 1 in 5 of these people ends up with another autoimmune glandular disease (autoimmune Addison's disease, Hasmimoto's autoimmune thyroiditis, Grave's disease of the thyroid). Likewise, plenty of people, with or without other autoantibodies, have anti-islet cell antibodies but never go on to develop autoimmune diabetes.

The claim from the early 1990's that cow's milk is the trigger for autoimmune diabetes flopped (JAMA 276: 609 & 647, 1996, NEJM 329: 1853, 1993, J. Clin. Endo. 87: 3192, 2002; the activists keep the business going but there now seems to be a consensus there is nothing real here: Eur. J. Clin. Nutr. 59: 623, 2005; Nutrition 21: 474, 2005).

* All the recent stuff is from obvious "independent thinkers" (ignoring what we know of immunology: Food & Chem. Tox. 42: 707, 2004) and studies that invited recall bias (Ann. Nutr. Metab. 47: 267, 2003).

* The non-obese diabetic mouse does get some protection from drinking mother's milk instead of cow's milk. The experimentalists speculat at length about how perhaps this is because mother's milk contains insulin and/or other peptides to which the gut lymphocytes need to become tolerant (Diabetes 48: 1501, 1999). But think -- the experiment requires taking the experimental mice away from their mothers. This must have many far-reaching effects beyond just the exposure to cow's milk.

A few groups are "curing" mice of type I diabetes using immune manipulation (Freund's adjuvants, etc., etc.) Update J. Clin. Inv. 108: 63, 2001.

Viral factors: Clinically, Type I diabetes often follows a viral illness.

Worth knowing: Kilham rat parvovirus infection produces type I autoimmune diabetes in diabetes-resistant rats (Diabetes 45: 557, 1996; J. Immuno. 165: 2866, 2000). This is now a robust finding (J. Imm. 173: 137, 2004).

* Retrovirus IDDMK(1,2)22 remains controversial as a cause of type I diabetes. Some folks don't find it at all (Diabetes 48: 209 & 219, 1999); others find soft data suggesting a link (J. Hum. Genet. 46: 712, 2001).

A Coxsackie B4 virus from the pancreas of a patient dying shortly after the onset of the illness destroys the beta cells of NOD (non-obese diabetic) mice; it's now clear that the virus causes a chronic infection of these islands (J. Inf. Dis. 171: 1131, 1995). Since this article, Coxsackie CB4 has been found commonly as a recent infection in kids coming down with diabetes.

In the non-obese diabetic (NOD) mouse, there is a specific defect in a group of T-suppressor cells. * A possible ligand to enhance these cells in mice and humans: Nat. Med. 7: 1052, 2001.

The mechanism of Coxsackie B4 induction of diabetes now seems clear -- the NOD mouse has lots of autoreactive-but-unactivated T-cells, Coxsackie B4 produces a mild infection of the beta cells, and bystander T-cells are activated ("bystander activation"). Happens in mice and maybe in kids.

* There have been outbreaks of type I diabetes in which the children were found to produce enterovirus mRNA; these patients do not show the usual HLA antigens cited as placing a person at risk, and perhaps this is a distinct subcategory with a specific viral cause (J. Inf. Dis. 187: 1562, 2003).

Overwhelming infections with mumps or cytomegalovirus also have been implicated in rare cases of "type I diabetes". The pancreas can be destroyed by congenital rubella.

* A huge search for the "insulitis virus" in humans using molecular probes found nothing: JAMA 257: 1145, 1987.

What does all this mean? In most cases of type I diabetes, it is hypothesized that a viral infection triggers autoimmune destruction of the beta cells in genetically-predisposed individuals.

However, in most of these children, there have been progressive abnormalities of glucose metabolism in these patients long before the onset of illness (Br. Med. J. 294: 5, 1987).

Some stress ("maybe the virus") apparently causes decompensation at the "time of onset". Following recovery from the first episode of ketoacidosis, the "honeymoon period" begins, when control is easy for several years. (* Patients continue to produce some of their own insulin -- i.e., there is C-peptide in their blood -- during the "honeymoon".)

* No one knows why type I diabetes is becoming more common, but there's no question that this is happening. Studies that suggest a major, unknown environmental factor (Lancet 364: 1699, 2004) make me think we are dealing with a viral trigger.