.Ketosis-prone diabetes: dissection of a heterogeneous syndrome using an immunogenetic and beta-cell functional classification, prospective analysis, and clinical outcomes
It is classified by antibodies and pancreatic capacity. This seems obscure but it isn't very hard.Analysis of clinical, phenotypic, and genotypic data derived from this prospective characterization of multiethnic, heterogeneous, ketosis-prone diabetic patients indicates the presence of novel forms of ß-cell dysfunction as well as a classification scheme to categorize these patients. We propose four groups based on two important features commonly used to distinguish type 1 and type 2 diabetes: presence or absence of biological markers of ß-cell autoimmunity, and presence or complete absence of ß-cell functional reserve. This is not meant to be rigid classification, but rather a hypothesis-testing scheme to differentiate etiologically and clinically distinct forms of ketosis-prone diabetic syndromes, and thus to uncover novel forms of ß-cell dysfunction. The distinctive pathogenetic features and diagnostic implications of the four Aß groups are discussed individually below.
Patients in this group, with significantly low ß-cell functional reserve together with circulating ß-cell autoantibodies, are likely identical with the well-defined form of autoimmune type 1 diabetes. They had early onset diabetes and were generally lean. African-American patients predominated in this group. The results of the HLA analysis supported the contention that these patients have typical autoimmune type 1 diabetes. Irrespective of ethnicity, certain HLA allelic variants are found in high frequency in persons with autoimmune type 1 diabetes (19, 24, 25, 26, 27, 28, 29, 30,31). The proportion of patients with the type 1 diabetes susceptibility HLA alleles DQB1*02 and DQA*03 was significantly higher in the A+ß- group than in the three other groups, including the phenotypically similar A-ß- group. Furthermore, no A+ß- patients were positive for the protective HLA alleles DRB1*15 and DQB1*0602 (19, 32, 33, 34, 35, 36). All patients in this group required multiple daily insulin injections to avoid ketosis 12 months after the episode of DKA, and a significant proportion had recurrence of DKA during this period despite close monitoring by the study team.
Patients in this group are likely to have diverse pathogenic mechanisms leading to ketosis-prone diabetes, including potentially novel forms of nonautoimmune ß-cell failure. There were numerous similarities in clinical characteristics and ß-cell functional reserve between the A+ß- and A-ß- groups (Table 2 and Figs. 2–4). At first glance, the difference between these two groups appeared to lie solely in their autoantibody status. However, HLA analysis revealed that there were also major differences between these two groups in genetic susceptibility to ß-cell autoimmunity. The frequencies of one class II allele (DQB1*02), which is strongly associated with autoimmune type 1 diabetes susceptibility (24, 29, 32, 37), and of another (DQA*03), which is in linkage disequilibrium with the strong susceptibility alleles DQB1*0302 and DQB1*0301, were low in the A-ß- group compared with the A+ß- group (Fig. 4). These features make it likely that the A-ß- group consists primarily of persons with nonautoimmune mechanisms of ß-cell injury, rather than persons with autoimmune type 1 diabetes whose circulating autoantibody levels have declined over time to undetectable levels (38). No A+ß- patients were positive for the protective allele DQB*0602 (33, 35, 39), whereas 9% of A-ß- patients possessed this allele. (There were no statistically significant group differences in the frequency of DQB*0602, however, probably because of the small sample sizes as well as the relatively low prevalence of the DQB*0602 allele in the general population (40). A-ß- patients also were more likely to have first-degree relatives with type 2 diabetes. The current classification scheme of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (41) would tend to place patients in the A-ß- group into the clinical category of idiopathic type 1 diabetes, a category that begs further definition, as provided by the criteria presented here.
Some patients in this group may represent a variant of what several reports of European cohorts have termed antibody-positive type 2 diabetes (42, 43, 44) or latent autoimmune diabetes ofadults (45, 46). However, others in the A+ß+ group likely represent a more aggressive form of late-onset autoimmune type 1 diabetes than described in these reports. DQB1*02 may be a marker for the more aggressive subset of A+ß+, because the six A+ß+ patients with DQB1*02 had higher mean HbA1c (8.6 ± 2.5%) than those without DQB1*02 (6.5 ± 0.6%) after 12 months of close management (P = 0.05). Furthermore, five of the six patients with DQB1*02 still require insulin treatment to avoid ketosis after 12 months of follow-up, whereas insulin has been discontinued safely in four of the five A+ß+ patients who lack this allele (P = 0.03). Although analysis of a larger cohort of A+ß+ patients is needed to confirm this suggestive trend, this combination of class II HLA and autoantibody markers may represent an important diagnostic opportunity to identify A+ß+ patients destined to have a more aggressive course. Because the presence of both the genetic markers and autoantibodies should precede the onset of clinical manifestations, it may be possible to identify such patients before their ß-cells are irreversibly destroyed (47).
This is the largest group of ketosis-prone patients, comprising the greatest number with new-onset diabetes. The frequencies of the autoimmune type 1 diabetes susceptibility HLA allelesDQB1*02 and DQA*03 are low in this group. A-ß+ patients appear clinically heterogeneous, with a wide range of BMI (Table 2). A-ß+ patients have achieved good glycemic control within 6 months of follow-up, and half have been able to discontinue insulin treatment.
The causes of severe, acute ß-cell dysfunction leading to DKA are likely to be diverse in this group. Half the A-ß+ patients have new-onset diabetes, without a notable precipitating factor for DKA. The mean HbA1c of this subgroup at presentation with DKA was 13.9 ± 2.2, indicating a relatively long period of undetected and untreated hyperglycemia. It is possible that the cause of acute ß-cell failure in these patients was glucotoxicity (48, 49, 50) or lipotoxicity (51), which reversed with excellent control of glycemia after the episode of DKA. The sustained, preserved ß-cell functional reserve and glycemic improvement in these patients argue against the likelihood that they have a form of type 1 diabetes with the poorly defined honeymoon period (52). In fact, all A-ß+ patients have now been evaluated for more than 1 yr, and one third for more than 2 yr, and they continue to maintain uniformly excellent glycemic control (mean HbA1c 7.0%) with adequatefasting levels of C-peptide (1.25 nmol/liter). The subset of A-ß+ patients with previously diagnosed diabetes may comprise patients with long-standing forms of type 2 diabetes with progressive ß-cell failure (53, 54) of such causes as ß-cell apoptosis (55), islet cell amyloid (56), or iron infiltration (57).
Three previous studies have measured islet cell autoantibodies and ß-cell function in subsets of African-American patients presenting with DKA (5, 6, 7). The patients described in these studies (e.g. those with "Flatbush diabetes") would fit into our two ß+ groups. Consistent with our ß+ group data, the mean age at diagnosis of these African-American cohorts was in the fifth decade, the mean BMI was high, only a minority had ß-cell autoantibodies, and glycemic control improved markedly after intensive treatment. These similarities add support to the concept of the A-ß+ group as manifesting a distinct form of ketosis-prone diabetes, but our data extend the expression of this syndrome to patients of Hispanic, Caucasian, and Asian ethnicity.
HLA genotyping was particularly helpful in distinguishing autoimmune-associated from probable nonautoimmune-associated forms of ß-cell dysfunction within the class of patients with low ß-cellfunctional reserve (i.e. in distinguishing the A+ß- and A-ß- syndromes). In the initial analysis, the class II alleles selected were those known to be strongly associated with autoimmune type 1 diabetes in multiple ethnic groups, e.g. the positively associated DQB1*02 and DQB1*0302 (22, 24, 27, 58, 59, 60, 61, 62, 63) and the negatively associated DQB1*0602 (19, 23, 32, 33,35, 39). In the pair-wise comparison, there was a clear difference in the relative frequencies of DQB1*02: high in the A+ß- group (72%) and low in the A-ß- group (26%). The frequency of DQB1*0302 showed a trend in the same direction, but did not attain significance after Bonferroni adjustment (which may not be necessary, because the association between this allele and autoimmune type 1 diabetes is well established). The protective allele DQB1*0602 (64) was absent in all patients in the A+ß- group, but present in 9% of A-ß- patients. DQB1*0602 is a low-frequency allele in the general population of Caucasian-Americans (5–13%) and African-Americans (4–15%) (40), hence a larger sample of patients would be necessary to have the power to detect group differences in its frequency. Interestingly, DQA*03, an allele not frequently reported to be associated per se with autoimmune type 1 diabetes susceptibility, also distinguished the A+ß- group (89%) from the A-ß- group (44%). DQA*03 is known to be in linkage disequilibrium with the strong susceptibility alleles DQB1*0302 and DQB1*0301, hence its frequency distribution is likely to represent a real difference in susceptibility to autoimmunity between the A+ß- and A-ß- groups.
The absence of features of autoimmune diabetes or HLA-associated susceptibility to autoimmune diabetes in the A- groups raises the possibility that they could include persons with geneticcauses of ß-cell dysfunction, such as syndromes of maturity onset diabetes of youth (MODY) or mitochondrial transfer RNA mutations. The MODY syndromes are characterized by Mendeliandominant inheritance due to monogenic mutations (65). Although there are at present no reported cases of subjects with documented MODY gene mutations presenting with ketoacidosis, this is certainly a possibility. Sixty-four (86%) of the patients in our A- cohort have a family history of type 2 diabetes, 45 of these with a potentially dominant mode of transmission. Screening of theextended pedigrees for linkage to the currently known MODY genes is ongoing. Diabetes associated with mitochondrial gene mutations also involves defects in glucose-stimulated insulin secretion (66). However, the absence of evidence for maternal transmission of diabetes and other typical features (e.g. deafness, neurologic disorders, cardiac or renal failure) make it unlikely that any of our patients harbor known mitochondrial gene mutations.
Imagawa et al. (67) have described a cohort of lean Japanese subjects who developed new-onset, fulminant ß-cell failure of apparently nonautoimmune cause after a relatively short period of hyperglycemia (HbA1c < 8%). Our two A- groups do not appear to include such patients, inasmuch as all of our A- patients, including those who were of new onset, had significantly higher HbA1c levels, a less fulminant course, greater BMI and higher frequency of first-degree relatives with diabetes. Furthermore, it is not clear that the Japanese patients were truly nonautoimmune,because they possessed HLA haplotypes (DRB1, DQA1, DQB1 0405,0303,0401, or DQB1 0901,0302,0303, or 0802,0401,0302) known to be associated with autoimmune type 1 diabetes (68, 69, 70).
The clinical course of the two ß- groups highlights the critical importance of ß-cell functional reserve in achieving effective glycemic control. Although both ß- groups experienced significant (3%) decreases in HbA1c and marked declines in the rate of hospital readmissions for DKA as a result of the study intervention, their chronic glycemic status remained quite poor. Other factors, such as lack of compliance with insulin treatment, could also have played a role in this outcome. We did not systematically record treatment compliance, but it is well-known that treatment noncompliance is particularly severe and glycemic control is especially difficult to achieve in type 1 diabetic patients in indigent, minority-ethnic, urban settings in the United States (42, 43, 44).
In conclusion, we have used a heterogeneous, multiethnic cohort to demonstrate that patients presenting with DKA comprise at least four distinct diabetic syndromes that are separable byautoantibody status, HLA genotype, and quantitative assessment of ß-cell function. Novel, nonautoimmune causes resulting in variable degrees of ß-cell dysfunction are likely to underlie the A-ß+ and A-ß- syndromes. Detailed genotypic and phenotypic characterization studies of patients in these categories are ongoing, in the hope that they will specify the etiologic bases of the syndromes revealed by the present analysis. The current data are also of clinical relevance to the evaluation and prognosis of patients with ketosis-prone diabetes. ß-Cell functional reserve at the time ofDKA is the strongest indicator of future metabolic control, but GAD and IA-2 autoantibody status and class II HLA allelotypes can assist in classifying ketosis-prone patients and improvingprediction of clinical outcomes.
Syndromes of Ketosis-Prone Diabetes Mellitus
Ketosis-prone diabetes (KPD) is a widespread, emerging, heterogeneous syndrome characterized by patients who present with diabetic ketoacidosis or unprovoked ketosis but do not necessarily have the typical phenotype of autoimmune type 1 diabetes. Multiple, severe forms of β-cell dysfunction appear to underlie the pathophysiology of KPD. Until recently, the syndrome has lacked an accurate, clinically relevant and etiologically useful classification scheme. We have utilized a large, longitudinally followed, heterogeneous, multiethnic cohort of KPD patients to identify four clinically and pathophysiologically distinct subgroups that are separable by the presence or absence of β-cell autoimmunity and the presence or absence of β-cell functional reserve. The resulting "Aβ" classification system of KPD has proven to be highly accurate and predictive of such clinically important outcomes as glycemic control and insulin dependence, as well as an aid to biochemical and molecular investigations into novel causes of β-cell dysfunction
|IV. Classification of KPD|
|TopAbstractI. IntroductionII. Case ReportsIII. History of KPDIV. Classification of KPD|
V. Natural History and...VI. Pathophysiology of KPD...VII. Management of KPDVIII. Conclusion and ProspectsReferences
To date, attempts to differentiate patients with KPD into clinically distinct and relevant subgroups have resulted in four different classification schemes: the ADA classification, a BMI-based system, a modified ADA classification, and the Aβ system.
The first is contained within the ADA’s most recent classification of diabetes in general (15) and has been adopted by investigators at the University of Texas Southwestern Medical School (Dallas, TX). All patients who experience DKA are defined as having type 1 diabetes, and among this group those who lack autoantibodies are referred to as "idiopathic type 1" or "type 1b." Strictly interpreted, the ADA scheme would define patients with both type 1a and type 1b diabetes as insulin dependent, because it does not mention possible reversion to insulin independence in either category; however, the Dallas group considers patients with type 1b to behave more like patients with type 2 diabetes, with some becoming insulin-independent. A second scheme is that developed by investigators at Emory University (Atlanta, GA) who separate KPD patients into lean or obese (9). "Lean KPD" patients are those with clinical characteristics of type 1 diabetes with low β-cell function, whereas "obese KPD" patients are those with clinical characteristics of type 2 diabetes with some preservation of β-cell function. A modification of the ADA scheme is used by investigators at the University of Paris who divide KPD patients into three groups (20). Patients with β-cell autoantibodies are classified as type 1a just as in the ADA scheme, whereas those who lack autoantibodies are distinguished retroactively, based on long-term insulin dependence, into "KPD insulin-dependent" (KPD-ID) and "KPD non-insulin dependent" (KPD-NID). Both type 1a and KPD-ID patients have clinical characteristics of type 1 diabetes with poor β-cell function, whereas subjects with KPD-NID have clinical characteristics of type 2 diabetes with preserved β-cell function for a prolonged duration.
Our collaborative group at Baylor College of Medicine and the University of Washington has used a classification system that distinguishes four KPD subgroups based on the presence or absence of autoantibodies and the presence or absence of β-cell functional reserve (Aβ classification) (1). The four subgroups are: A+β– (patients with autoantibodies and absent β-cell function); A+β+ (those with autoantibodies but preserved β-cell functional reserve); A–β– (those without autoantibodies but absent β-cell function); and A–β+ (those without autoantibodies and preservedβ-cell functional reserve). A+β– and A–β– patients are immunologically and genetically distinct from each other but share clinical characteristics of type 1 diabetes with very low β-cell function, whereas A+β+ and A–β+ patients are immunologically and genetically distinct from each other but share clinical characteristics of type 2 diabetes with preserved β-cell functional reserve (Fig. 1and
There will be a test on this later so please read up.