Post mortem examination of the pancreas in long-term diabetes shows an absence or greatly reduced number of insulin-staining beta cells; the residual pseudoatrophic islets contain abundant alpha cells and show light fibrosis. Our knowledge of earlier abnormalities is limited by the rarity of pancreatic specimens from around the time of diagnosis, which have mostly been obtained from children who died in ketoacidosis. A proportion show the characteristic appearance of insulitis, due to lymphocytic infiltration of the islets. This transient lesion is most commonly observed in the young and in those with the most recent onset. A search is under way for histological abnormalities in non-diabetic individuals at high risk of progression to diabetes, and lymphocytic infiltration has been observed in rare instances. There is a high level of scientific interest in the nature and cause of the earliest lesions associated with progression to type 1 diabetes, since this might hold important clues as to its aetiology. The specificity of the disease process for the beta cell (as against other islet cell types) is still not well understood, and the potential for beta cell regeneration in response to immune attack has also generated intense interest.
The pancreatic islets are difficult territory for the pathologist, since they represent no more than 1–2% of the total weight of the gland, and the pattern of islet damage may be quite variable within the same individual. Furthermore, the organ suffers from autodigestion soon after death due to the action of enzymes within the exocrine pancreas, and fresh material may be hard to obtain. Early pathologists were puzzled by the relatively normal appearance of the pancreas under the microscope in newly diagnosed diabetes, and special stains and techniques were needed to pinpoint abnormalities. The hallmark change associated with recent-onset juvenile diabetes is round-celled (lymphocytic) infiltration of the islets, referred to as insulitis. This was first observed in 1902, but its potential significance was unknown at the time, since the role and function of the lymphocyte was not understood until the 1950s. The classic description of insulitis was provided by Willy Geptsin 1965, who speculated that this provided evidence for an immune basis for juvenile diabetes. Insulitis is also present in experimental models such as the NOD mouse, but differs in some features, and extrapolation from mouse to man should therefore be undertaken with caution.
This article and subsequent sections summarise the main changes seen in the pancreas in long-standing and recent onset diabetes, and in individuals at high risk of progression to type 1 diabetes. Possible reasons for beta cell vulnerability to immune attack are considered, together with evidence for attempted beta cell regeneration.
The pancreas in long-term diabetes
Type 1 diabetes of long duration is associated with loss of pancreatic weight, due to atrophy of exocrine tissue. This is believed to result from loss of the trophic effects of locally produced insulin upon the surrounding acinar tissue. The majority of insulin-staining beta cells have disappeared, and the resulting pseudo-atrophic islets are characterised by abundant glucagon-secreting alpha cells and minor fibrosis of the islet interstitium. Up to 50% of patients may however retain some residual beta cells, even after 50 years of diabetes. 
The pancreas in recent-onset diabetes
Death following the onset of type 1 diabetes is mercifully rare, and relatively few fresh pancreatic specimens have been available for up-to-date forms of investigation. Insulitis is a notably elusive lesion, and has no generally accepted definition. At diagnosis different regions of the same pancreas may show apparently healthy islets, islets invaded by lymphocytes and islets containing no beta cells at all; such differences typically show a lobular distribution. When present, insulitis typically affects 8–10% of islets. Although the lesion is considered typical of type 1 diabetes, only about 150 examples have been described in the literature. It has been found in 75% of children diagnosed under the age of 14 years who died within a month of diagnosis, in 60% who died within 1–12 months of diagnosis, and in only 4% with a disease duration >12 months. In contrast, insulitis was found in only 30% of those aged 15–40 years who died within 1 month of diagnosis. The inverse association between insulitis and age at onset of diabetes might be due to a more fulminant onset of disease, but is otherwise unexplained. Beta cell replication appears to be stimulated by inflammatory changes in the islets, and it is possible that this in turn promotes further immune attack. 
It has in some instances been possible to characterise the infiltrate, which has been noted to show some variation from case to case. In general, CD8+ cytotoxic/suppressor cells predominate in the early stages, followed by CD20+ B cells; both cell types tend to disappear in parallel with the insulin content of the pancreas. In contrast, macrophages sometimes predominate and may persist in insulin-deficient islets1.
Insulitis before the onset of diabetes
Islet autoantibodies can be detected years before the clinical onset of diabetes, and this has prompted a search for the earliest detectable histological change in type 1 diabetes. This has been achieved by screening organ donors for islet autoantibodies and examining the pancreas in those who test positive. In one study of antibody-positive donors, insulitis was found in 2/62 pancreases; both were obtained from donors with multiple islet autoantibodies, each of whom had a relatively normal beta cell mass.5
Do beta cells replicate?
This is one of the most important and controversial issues in the study of diabetes. The existence of a mechanism for beta cell replication might potentially explain observed differences in the rate of progression to overt diabetes in antibody-positive individuals, and might – above all – open up new and exciting therapeutic possibilities. The possibility of successful replication is supported by differences in the beta cell mass of longer duration patients, 30–50% of whom may retain small numbers of functional insulin-secreting cells for more than 20 years after diagnosis.
It is not known whether persistence of beta cells in the islets represents simple survival of existing beta cells or new cell formation. It is therefore of interest that nine patients with a diabetes duration from 52–84 years all showed insulin-staining cells in their pancreases.4 A lobular pattern of distribution was described in the same patients, with evidence of beta cell regeneration and apoptosis. A separate study of long-duration patients found that 6/20 had residual beta cells; some had a mix of insulin-staining and insulin-deficient islets (pattern A), whereas others had some insulin-containing cells in all islets (pattern B).
Two recent studies have provided tentative evidence of a 'futile cycle' of new beta cell formation and apoptosis in chronic type 1 diabetes,  and this observation clearly merits further investigation.
^ In't Veld P. Insulitis in human type 1 diabetes. The quest for an elusive lesion. Islets 2011;4:131–8
^ Gepts W. Pathologic anatomy of the pancreas in juvenile diabetes mellitus. Diabetes 1965;14:619–33
^ Lohr M, Kloppel G. Residual insulin positivity and pancreatic atrophy in relation to duration of chronic type 1 diabetes mellitus and microangiopathy. Diabetologia 1987;30:757–62
^ Keenan HA et al. Residual insulin production and pancreatic turnover after 50 years of Joslin Medalist Study. Diabetes 2010;59:2846–53
^ In't Veld P et al. Screening for insulitis in adult autoantibody positive organ donors. Diabetes 2007;56:2400–4
^ Willcox A et al. Evidence of increased islet cell proliferation in patients with recent-onset type 1 diabetes. Diabetologia 2010;53:2020–8
^ Gianani R et al. Dimorphic histopathology of long-standing childhood-onset diabetes. Diabetologia 2010;53:690–8
^ Meier JJ et al. Sustained beta cell apoptosis in patients with long-standing type 1 indirect evidence of islet regeneration? Diabetologia 2005;48:2221–9
^ Keenan HA et al. Residual insulin production and pancreatic beta cell turnover after 50 years of Joslin Medalist Study. Diabetes 2010;59:2846–53