Prevalence and impact of initial misclassification of pediatric type 1 diabetes mellitus

From the Department of Epidemiology and Biostatistics, University of South Carolina Arnold School of Public Health, and the Departments of Internal Medicine, Pediatrics, and Neuropsychiatry and Behavioral Science, University of South Carolina School of Medicine, Columbia.
Southern medical journal (Impact Factor: 0.93). 10/2012; 105(10):513-7. DOI: 10.1097/SMJ.0b013e318268ca60
Source: PubMed


To characterize rates of initial misclassification of type 1 diabetes mellitus as type 2/unspecified diabetes mellitus in a cohort of children/adolescents and to examine the impact of misclassification on the risk of diabetes-related complications.
An 11-year dataset (1996-2006) was analyzed. Inclusion criteria included age 17 years and younger, enrollees in South Carolina State Medicaid, and diagnosis of type 2/unspecified or type 1 diabetes mellitus for at least two visits, 15 days apart. Survival analysis was used to assess the association of "misclassification" with the incidence of diabetic ketoacidosis (DKA), and the cumulative incidence of neuropathy, nephropathy, and cardiovascular complications, after controlling for individual risk factors and comorbid conditions.
A total of 1130 individuals meeting the inclusion criteria were studied for a median of 7 years. Of the 1130 individuals, 669 (59.2%) maintained a diagnosis of type 2/unspecified diabetes mellitus, 205 (18.1%) were consistently diagnosed as type 1 diabetes mellitus, and the remaining 256 individuals (22.7%) were misclassified. Insulin treatment was used in 100% of the type 1 diabetes mellitus group and 73% of the misclassified group. Compared with the type 2 diabetes mellitus group, being misclassified was associated with earlier development of DKA (adjusted hazard ratio [aHR] 5.08, 95% confidence interval [CI] 3.09-8.37), neuropathy (aHR 1.94, CI 1.31-2.88), and nephropathy (aHR 1.72, CI 1.19-2.50), whereas being consistently classified with type 1 diabetes mellitus was associated only with earlier development of DKA (aHR 4.96, CI 2.56-9.61).
Proper categorization of pediatric diabetes can be challenging, especially with comorbid obesity. Failure to ascertain type 1 diabetes mellitus in a timely manner in a pediatric population may increase the risk of substandard care and diabetes-related complications.

Download full-text


Available from: Jeanette M Jerrell, Oct 05, 2015
1 Follower
73 Reads
  • Southern medical journal 10/2012; 105(10):518-9. DOI:10.1097/SMJ.0b013e318268cb81 · 0.93 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVE To evaluate whether etiologic diabetes type is associated with the degree of albuminuria in children with diabetes.RESEARCH DESIGN AND METHODSSEARCH is an observational, longitudinal study of children with diabetes. Youth with newly diagnosed diabetes were classified according to diabetes autoantibody (DAA) status and presence of insulin resistance. We defined insulin resistance as an insulin sensitivity score <25th percentile for the United States general youth population. DAA status was based on positivity for the 65-kD isoform of glutamate decarboxylase and insulinoma-associated protein 2 antigens. The four etiologic diabetes type groups were as follows: DAA(+)/insulin-sensitive (IS) (n = 1,351); DAA(+)/insulin-resistant (IR) (n = 438); DAA(-)/IR (n = 379); and DAA(-)/IS (n = 233). Urinary albumin:creatinine ratio (UACR) was measured from a random urine specimen. Multivariable regression analyses assessed the independent relationship between the four diabetes type groups and magnitude of UACR.RESULTSAdjusted UACR means across the four groups were as follows: DAA(+)/IS = 154 μg/mg; DAA(+)/IR = 137 μg/mg; DAA(-)/IR = 257 μg/mg; and DAA(-)/IS = 131 μg/mg (P < 0.005). Only DAA(-)/IR was significantly different. We performed post hoc multivariable regression analysis restricted to the two IR groups to explore the contribution of DAA status and insulin sensitivity (continuous) to the difference in UACR between the IR groups. Only insulin sensitivity was significantly associated with UACR (β = -0.54; P < 0.0001).CONCLUSIONS In youth with diabetes, the DAA(-)/IR group had a greater UACR than all other groups, possibly because of the greater magnitude of insulin resistance. Further exploration of the relationships between severity of insulin resistance, autoimmunity, and albuminuria in youth with diabetes is warranted.
    Diabetes care 07/2013; 36(11). DOI:10.2337/dc13-0568 · 8.42 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: National data of type 1 diabetes incidence are currently missing in Italy. To fill this gap, we estimated the national rate of first hospitalizations for type 1 diabetes among children aged 0-4 years and resident in Italy, as well as rates for each of the twenty-one Italian regions. We extracted the first episode of hospitalization in years 2005-2010 from the Italian Hospital Discharge Registry (IHDR). Record-linkage procedure and cleansing data method were applied to exclude prevalent cases and potentially miscoded patients. At the end, 2,250 incident hospitalizations for type 1 diabetes were extracted. In the years 2005-2010, the mean nation-wide first hospitalization rate for type 1 diabetes in children aged 0-4 years was 13.4 (95 % CI 12.8-14.0), 14.1 (95 % CI 13.3-14.9) in males and 12.7 (95 % CI 11.9-13.4) in females. A geographically heterogeneous pattern of incidence was found: even excluding Sardinia, incidence for this age range and calendar period tended to be slightly higher in Southern than in Northern Italy. Our incidence of first hospitalizations corresponds to the estimates of disease incidence obtained with different data sources by other authors in selected Italian regions. We provide, for the first time ever, the estimate of type 1 diabetes incidence for the overall population aged 0-4 years resident in Italy. When methodological cautions are adopted, IHDR emerges as a reasonable proxy of type 1 diabetes incidence and as a cost-effective tool for public health purposes.
    Acta Diabetologica 01/2014; 51(2). DOI:10.1007/s00592-014-0556-4 · 2.40 Impact Factor
Show more