Expansion in health coverage
In presenting the findings of a study showing an increase in the prevalence of DKA at diagnosis of type 1 diabetes in children in Colorado from 2010 to 2017, Dr. Rewers said that the increase “paradoxically occurred” at a time of increasing health insurance coverage, a reference to the expansion of Medicaid during 2008-2012 and implementation in 2013 of the Affordable Care Act.
“Our group in Colorado has followed the frequency of DKA for almost 2 decades,” Dr. Rewers said. It’s important to study DKA as it is linked to worse glycemic control – with children with DKA having an HbA1c level of around 1% higher than those without DKA – and the potential for future, long-term complications.
Dr. Rewers noted that the increase in DKA at diagnosis of type 1 diabetes was more rapid in the children who had private rather than public health insurance. Of 1,187 patients with DKA, 57% had private health insurance, and 37% had public insurance, compared with 66% and 28%, respectively, in those without DKA. In 2010, the prevalence of DKA at diagnosis was 35.3% in those who were privately insured and 52.2% of those with public health insurance, but by 2017, a similar percentage of DKA was seen in the privately and publicly insured children (59.6% and 58.5%, respectively).
She said one possible explanation for that might be that “increased enrollment in high-deductible insurance plans could discourage families with private insurance from seeking timely care.”
Another explanation is that there is a low awareness of type 1 diabetes in the general population, she added. “Educational campaigns and autoimmunity screening have been shown to reduce DKA at diabetes diagnosis, but unfortunately they are not used widely at this point.”
Identifying at-risk children
“Diabetic ketoacidosis is a serious complication of diabetes [and] is difficult to diagnose because of the variability of the symptoms, said Angela Ibald-Mulli, PhD, who presented the findings of a retrospective cohort study in which she and her colleagues used a “discovery algorithm” called Q-Finder to identify the predictive factors for DKA in youth with type 1 diabetes, based on data from the Diabetes Prospective Follow-up Registry (DPV).
Sara Freeman/MDedge News
Dr. Angela Ibald-Mulli
“The better we know the risk factors, the better we can care for our patients,” she emphasized.
The investigators obtained data on 108,223 patients with a diagnosis of type 1 disease and with more than two visits related to diabetes. The prevalence of DKA – defined as a pH of less than 7.3 during hospitalization occurring at least 10 days after the onset of type 1 diabetes – was 5.2%, said Dr. Ibald-Mulli, head of Medical Evidence Generation Primary Care at Sanofi, Paris.
A total of 129 different features were considered for their association with DKA – including comorbidities, sociodemographic factors, laboratory values, and concomitant medications – and were then used to identify, test, and the validate likely risk profiles.
After comparing the characteristics of patients with and without DKA, eight significant factors, all of which have been reported previously in the DPV cohort, were seen: younger age, lower body weight, higher HbA1c, younger age at onset of T1D; shorter disease duration; having a migration background; being less active; and having had more medical visits.
The investigators used the algorithm, and found 11 distinct profiles associated with DKA: an HbA1c higher than 8.87%; being aged 6-10 years; being aged 11-15 years; a diagnosis of nephropathy; DKA being present at onset; a prevalence of hypoglycemia with coma; a diagnosis of thyroiditis; a standardized body mass index lower than 16.9; not using short-acting insulin; younger than age 15 years; and not using continuous glucose monitoring.
Almost two-thirds of patients (64.7%) belonged to at least one of these risk profiles, Dr. Ibald-Mulli observed, with 7.1% of them having DKA, compared with 1.6% who belonged to none of the profiles.
Dr. Ibald-Mulli said it was important to note that the DKA risk profiles could overlap. “The more profiles a patient belongs to, the higher is the risk of having DKA,” she emphasized, adding that most patients (88.8%) with DKA belonged to just one profile, and fewer than 5% belonged to three or more profiles.
“Overall, the results of the algorithm confirmed known risk-factor profiles that had been previously identified by conventional statistical methods,” she concluded. It also provided “additional insights that can be further explored.”
SEARCH is funded by the Centers for Disease and Prevention and the National Institute of Diabetes and Digestive and Kidney Diseases. The DPV Registry is funded by multiple sponsors, including the European Federation for the Study of Diabetes and other academic institutions with the support of several commercial partners. Sanofi sponsored the study presented by Dr. Ibald-Mulli. Dr. Rewers made no disclosures, and Dr. Jensen did not have any conflicts of interest to declare. Dr. Ibald-Mulli is an employee of Sanofi.
SOURCE: Rewers A et al. EASD 2019, Abstract 115; Jensen E et al. EASD 2019, Abstract 116; Ibald-Mulli A et al. EASD 2019, Abstract 117.