eLetters to:

ELECTRONIC ARTICLE: Eba H. Hathout, Wendy Thomas, Mohamed El-Shahawy, Fadi Nahab, and John W. Mace Diabetic Autoimmune Markers in Children and Adolescents With Type 2 Diabetes Pediatrics 2001; 107: e102 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

eLetters published:

Confusing risk with diagnosis leads to erroneous conclusions

Philip Zeitler, Orit Pinhas-Hamiel   (14 June 2001)

Untitled

Eba H Hathout   (6 July 2001)

Confusing risk with diagnosis leads to erroneous conclusions 14 June 2001
Philip Zeitler, Associate Professor of Pediatrics University of Colorado Health Science Center and Maccabi Diabetes Center, Israel, Orit Pinhas-Hamiel Send letter to journal: Re: Confusing risk with diagnosis leads to erroneous conclusions phil.zeitler{at}uchsc.edu Philip Zeitler, et al.	Dear Editor: We have read carefully the recent article by Drs Hathout et al (Diabetic autoimmunity markers in children and adolescents with type 2 diabetes, Pediatrics 107:E102, 2001). Though the article is intriguing, we are concerned that the conclusions proposed by the authors are based on an erroneous approach to patient classification. We are even more concerned that the conclusions drawn from this article will only serve to further confuse clinicians and complicate further an already complex area of research. According to the authors, type 2 diabetes was diagnosed when a patient met criteria for diabetes according to the ADA criteria and possessed two or more risk factors for type 2 diabetes. Classification of diabetes type was “confirmed” by measurement of basal and/or stimulated C- peptide. Finally, the authors measured islet cell autoantibodies and report that patients with type 2 have higher levels of autoantibody positivity than previously assumed. The authors conclusion that patients with type 2 diabetes can have autoantibodies present and that, therefore, autoantibodies are not reliable for the distinction of type 1 vs type 2 diabetes, depends entirely on the validity of the criteria used to classify patients into “Type 1” and “Type 2” groups in the first place. However, we take very strong exception to the diagnostic criteria applied and would suggest that it can be equally argued that patients in the “type 2” group with antibody positivity were mis-classified in the first place. First of all, there is nothing in the criteria used for classifying patients as Type 2 that is truly diagnostic. Each of these are simply risk factors. The fact of the matter is that obese, minority patients with family histories of type 2 diabetes still have a very significant probability of having type 1 rather than type 2 given the relative prevalence of the two disorders. Second, though the authors state that C- peptide was used to confirm the diagnosis of Type 2, the difference in C- peptide level between type 1 and type 2 patients is non-significant, making it unclear how this measurement was used as a distinguishing factor. The authors do not make it clear whether the C-peptide values presented in Table 2 are stimulated, basal, or some combination. Third, in the absence of presentation of data comparing antibody positive vs antibody negative “type 2” patients, it is not possible to evaluate whether there are differences between these groups in terms of clinical characteristics. We would suggest that the antibody positive “type 2” patients might be more accurately diagnosed as type 1 and segregation of the two groups would demonstrate whether insulin requirement was greater in one group than the other, whether there was a difference in the risk factors that lead to characterization as type 2 etc. The authors refer to studies in adult patients indicating that adult patients with Type 2 who have antibody positivity have a greater chance for development of insulin requirement over time. An alternative way to interpret these data is that type 1 diabetes is more common among adults than has been historically assumed and that these adult patients are, in fact, type 1, not type 2 patients with slowly developing disease and late onset of insulin requirement. This view is rapidly gaining currency among students of adult diabetes. Thus, we agree with the authors that the antibody positive type 2 adolescent has a lot in common with the antibody positive adult “type 2” patient, but we argue that both of these patients are in fact Type 1 patients in the sense that autoimmunity underlies the progression of their disorder. The bottom line is that using clinical characteristics to define type 2 diabetes and then demonstrating that laboratory measurements do not correlate well with this clinical definition does not necessarily mean that the laboratory measurement is the one in error. It is equally, if not more, likely that the initial clinical characterization is in error. The ultimate etiologic distinction between type 1 and type 2 diabetes relies on the demonstration of the presence and/or absence of insulin deficiency and degree of insulin resistance. Since there is no rigorous determination of the true etiology of diabetes from a physiologic perspective in the two groups, the initial classification is suspect. Therefore, conclusions about autoimmunity as a distinguishing feature between the two groups are not convincing. It is as reasonable to conclude that the initial classification is erroneous as it is to conclude that the initial classification is correct and that measurement of autoantibodies is unreliable.
Untitled 6 July 2001
Eba H Hathout, MD Pediatric Diabetes Center, Loma Linda University Children's Hospital Send letter to journal: Re: this article EHathout{at}ahs.llumc.edu Eba H Hathout	Thank you for your thorough review and critique of our article entitled “Diabetic Autoimmune Markers in Children and Adolescents with Type 2 Diabetes.” The letter raises a legitimate concern about accuracy of classification of the type of diabetes. For clarification, we would like to mention that our reported fasting C-peptide data was done with the majority of patients receiving insulin therapy. Hence, the lack of a significant difference in C-peptide levels between type 1 and type 2 patients cannot be interpreted as an error in diabetes classification. Basal and stimulated C-peptide measurement following insulin withdrawal was not done in type 1 patients. Therefore, comparative endogenous insulin production data for type 1 and type 2 patients could not be ascertained, and would require, as the commentary suggests, more elaborate islet function testing. Nevertheless, when patients classified as having type 2 diabetes were analyzed, there was no significant difference in age, BMI, C-peptide, HbA1c, or frequency of acanthosis nigricans or DKA at diagnosis between antibody-positive and antibody-negative patients. We would also like to bring up the following points. Firstly, it is a well-known, and hitherto undisputed fact, that a certain percentage of the general adult population has positive diabetes antibodies, without any evidence of either type of diabetes. Hence, it is not unreasonable to assume that a certain percentage of adolescents, or adults, with type 2 diabetes will have one or more diabetes antibody markers. There is no data to suggest that type 2 diabetes prevents or arrests the development of naturally-occurring diabetes or other antibodies. An analysis of the difference in the incidence of these markers between healthy and diabetic individuals would be helpful, although the natural history of diabetes antibodies is yet to be clearly outlined. Secondly, the argument that obese minority adolescents with a family history of diabetes and positive diabetes antibodies are more likely to have type 1 diabetes based on “relative prevalence” may be true of certain geographic locations and ethnic constituencies. However, it is certainly not the case based on current trends in our center in Southern California, and other minority-rich pediatric diabetes centers in the US. Our article also addressed the possibility that some adolescents may have both types of diabetes (double diabetes). Thirdly, in view of the acceptance of, and multiple publications on, antibody-negative type 1 diabetes, classification criteria of the type of diabetes cannot be rigidly established based on serologic data. Fourthly, diagnosis of type 2 diabetes at our center was certainly not arrived at based on historic information alone. At presentation, all patients had antibody testing, and fasting C-peptide measurements. The reported C-peptide data was only that obtained at initial presentation with patients on insulin therapy (which is probably the explanation for the insignificant difference between type 1 and type 2 subgroups). Stimulated C-peptide levels were not included because they were done at multiple stages of the diagnostic process to avoid erroneous interpretation due to either glucose toxicity or a honeymoon period. In many instances, the diagnostic process lasted one to two years, and in one instance it lead to the identification of the first Mexican family with genetically-documented MODY (work-up done at Dr. Graeme Bell’s laboratory, see our letter in Diabetes Care). In conclusion, larger multi-center studies, especially from US pediatric diabetes centers with expansive minority populations, are needed to confirm or refute our findings. Until then, we believe that the pediatric diabetes community should question therapeutic restrictions based on antibody criteria, when such restrictions are exclusively applied to children and not adults. Eba H. Hathout, MD John W. Mace, MD Pediatric Diabetes Center Loma Linda University Children’s Hospital Loma Linda, California