Correctly distinguishing the etiology of childhood diabetes has been an issue for many investigators. An incidence study of diabetes among Swedish youth aged 15 -34 demonstrated that even in this relatively homogeneous population with few structural barriers to diagnosis and optimal treatment, confusion as to the clinical type and etiology of diabetes can occur (42). Patients diagnosed in 1983-84 (n = 281) were followed for 3 years. Initially, 75% were classified as Type 1, 19% were Type 2, and 6% were unclassifiable or their diabetes was secondary to another disease process. By 3 years duration, 87% of the Type 1 diabetes patients were still classified as Type 1, and 72% of the initial Type 2 diabetes patients were still in that category. Thus, 13% of Type 1 diabetes and 28% of Type 2 patients (n = 43 in all) exhibited an atypical clinical course. Of these, six patients were designated Type 1 at onset on the basis of glycemia, ketonuria and other clinical characteristics. At followup, these patients remained lean, but had come off insulin without developing ketosis, and had C-peptide levels intermediate between those of the 'true' Type 1s and the Type 1 diabetic patients. Thus, even in a carefully defined population, correctly distinguishing Type 1 from atypical or early-onset Type 2 diabetes can be difficult based on clinical data alone.
A few investigators have hypothesized that Type 2 diabetes can coexist with insulin-deficient Type 1 diabetes, where insulin resistance develops with increasing insulin dosage. These 'double diabetes' patients are most often those with Type 2 relatives (43). In their longitudinal study of the natural history of chronic complications (n = 658 childhood onset Type 1 diabetes patients), Erbey et al. reported that those with incident coronary artery disease or overt nephropathy had higher levels of insulin resistance and more relatives with Type 2 diabetes (44).
Further confusion arises when certain data on adults with diabetes are considered. Correct definition is an absolute requirement for the epide-miologic investigation of a disease, yet 5-10% of most patients reported in clinic series reports have been unclassifiable into either major category of diabetes. This difficulty has been a topic of discussion for many years (42,45). Fasting and glucagon-stimulated C-peptide levels have been shown to distinguish Type 1 from Type 2 diabetes in all but a handful of European patients (46), although few long-term population-based data are available. Atypical diabetes among non-Europeans may be more akin to Type 2 diabetes with respect to insulin secretion, although at onset most atypical patients are clearly insulinopenic.
In adult blacks, Banerji et al. (47) have described a syndrome they refer to as 'Flatbush' diabetes. The 21 middle-aged African American patients in this report developed DKA, all but 4 of them at the time of diagnosis. Subsequently, however, the clinical course of their disease resembled Type 2 diabetes. At the time of the study, 3-120 months after the episodes of DKA, these patients were lean to mildly obese, and fewer than half were being treated with insulin. Diabetes was reported in a first-degree relative of 14 (67%) of the patients. Assays for ICA and GAD antibodies were negative, but an elevated frequency of the Type 1-related HLA alleles DR3 and DR4 was reported. Compared with non-diabetic adults (13 blacks, 3 whites), these patients had significantly lower insulin and higher glucose areas on 2-hour oral glucose tolerance testing, yet clamp studies revealed insulin resistance in all but one patient. The investigators concluded that the etiology in these adult patients must be a hybrid of Type 1 and Type 2 diabetes.
Banerji and colleagues conducted additional studies in blacks with Type 2 diabetes contrasting insulin-sensitive vs. insulin-resistant syndromes using the euglycemic clamp technique (48). Differences in visceral fat deposition were reported (49), and varying associations with HLA-DQ polymorphisms were identified among 25 insulinsensitive, 21 insulin-resistant, and 89 normogly-cemic African Americans. The insulin-resistant group had a higher than expected prevalence of the DQw7 allele, and the insulin-sensitive group had a lower frequency of DQw6, when compared with each other and with non-diabetic blacks. No HLA-DR or Class I differences were observed. Joffe et al. (2) reported data from South African black Type 2 diabetes patients suggesting a rapid fall in beta-cell function compared with whites, despite obesity and a relatively mild clinical course. There was no increase in C-peptide after sulfonyurea treatment for hyperglycemia. The beta-cell response to a glucose challenge was reduced in obese non-diabetic blacks when compared with obese non-diabetic whites. The authors cite these data as evidence in support of a primary defect in insulin production rather than in insulin sensitivity, perhaps related to childhood under-nutrition.
Was this article helpful?