What’s the relationship between COVID-19 and diabetes?
May saw the e-publication of the DIMPLES study. Let’s take a deep dive into this paper and its implications for PEM internationally.
Ponmani C, Nijman RG, Roland D, Barrett M, Hulse T, Whittle V, Lyttle MD on behalf of PERUKI. Children presenting with diabetes and diabetic ketoacidosis to Emergency Departments during the COVID-19 pandemic in the UK and Ireland: an international retrospective observational study. Arch Dis Child. 2023 May 17:archdischild-2022-325280. doi: 10.1136/archdischild-2022-325280.
Why was the DIMPLES study needed?
During the COVID-19 pandemic, we noticed a change in the type of patients presenting to the Emergency Department.
Despite a notable decline in overall presentations, anecdotally, it felt like there was an increase in the number of children presenting with new onset diabetes, diabetic ketoacidosis (DKA), and especially severe DKA.
Organisational changes across the PEM landscape made it difficult to establish a clear pattern. The European REPEM EPISODES study suggested that the relative increase in DKA cases may just have been due to our biased perception of an increased occurrence.
The DIMPLES study by PERUKI sought to investigate the incidence of new-onset diabetes in the paediatric population across the UK and Ireland, as well as the characteristics of these ED presentations, and to determine if SARS-CoV2 infection was involved.
Which patients were involved?
Children between the ages of 6 months to 16 years, across 49 paediatric EDs in the UK and Ireland who were member sites of Paediatric Emergency Research in the UK & Ireland (PERUKI), presenting with either
- New onset diabetes or
- Pre-existing diabetes with diabetic ketoacidosis (DKA)
They looked at two time points:
- During the COVID-19 pandemic (1st March 2020–28th February 2021)
- and the preceding year (1st March 2019–28th February 2020).
For the purpose of the study, DKA was defined as:
– pH < 7.3
– HCO3 <15mmol
– Ketones >3.0mmol/L
Severe DKA was defined as:
– pH < 7.1
– HCO3 <5mmol
– Ketones >3.0mmol/L
Children were identified by retrospectively searching electronic patient record data using ICD10 coding systems (10th revision codes for Diabetes and DKA were used). Identified cases were then screened further by looking at the medical notes to confirm whether the presentation was eligible for inclusion.
All ED visits that met the criteria for ‘new onset diabetes’ or DKA (‘decompensation’) were included.
What outcomes did they look at?
The primary aim of the study was to describe the incidence, clinical and biochemical characteristics of young people presenting to the ED with diabetes and to investigate whether there was a relationship to SARS-CoV-2.
The Secondary aim was to describe the severity and incidence of new presentations of DKA and decompensation of existing diabetes, compared to pre-COVID presentations.
What were the results of the DIMPLES study?
- New onset diabetes increased from 1015 to 1183 – a 17% increase – compared with a background incidence of 3-5% over the previous five years in the UK.
- Severe DKA pH <7.1 increased from 141 to 252 – a 79% rise (p<0.001)
- Admissions to PICU increased from 38 to 72 – an 89% rise (p<0.05)
- DKA pH < 7.3 increased from 395 to 566 – a 43% rise
- Children requiring fluid boluses of 20ml/kg or more in DKA in ED increased from 64 to 161 (p<0.001)
- The severity of new-onset diabetes was also reflected by other clinical and biochemical markers such as pH, degree of tachycardia and tachypnoea (all <0.001)
- The time from onset of symptoms to ED presentation was similar across the two study periods, suggesting that delay in presentation was not the sole contributing factor at play.
- The usual seasonal pattern of presentation during the COVID-19 pandemic was different. Seasonal variation appeared to disappear – diabetes presentations usually peak in winter with a trough in summer. This did not happen during the COVID-19 study period.
- Fewer children with pre-existing diabetes presented during the COVID-19 pandemic year than the year prior, and they had fewer episodes of decompensation.
How good was this paper? The CASP Checklist
Did the study address a clearly focused issue?
Yes. This was a well-defined population with a subsequent search of electronic patient records and chart reviews. Inclusion was based on strict criteria and internationally recognised standards/definitions of DKA.
Was the cohort recruited in an acceptable way?
Yes. The patients were identified through site leads from the PERUKI research network. Presentations meeting the ICD10 diagnosis of diabetes / DKA were identified and selected from electronic patient records by the leads. They then carried out a review of medical notes to ensure the criteria were met. All individual presentations recorded as DKA or diabetes (decompensation) were included.
Was the exposure accurately measured to minimise bias?
Some selection bias existed regarding testing for SARS-CoV-2 infection. Children were only being swabbed if they were symptomatic. Given the number of children with SARS-CoV-2 infection who showed very mild or, in many cases, even no symptoms, his is likely to be significant.
Antibody testing was not readily available, so prior infection with SARS-CoV-2, particularly in cases where the children remained asymptomatic, was unlikely to be known and could not be accurately recorded.
Was the outcome accurately measured to minimise bias?
All patients presenting with DKA or existing diabetes with decompensation who were captured on the electronic patient record using ICD10 criteria were included. This meant 2746 ED presentations by 2618 individuals.
Each case was examined carefully through chart review, and strict diagnostic criteria, consistent with internationally recognised definitions of DKA, were used. This helped reduce bias. However, only patients presenting to the ED were included. Although unlikely, this may have excluded some patients presenting elsewhere.
Have the authors identified all important confounding factors?
The authors referred to a potential presentation delay – expected during the COVID-19 pandemic. However, over the study periods, there was no significant delay in presentations to an emergency department. The previously reported seasonal variation in presentations was lost during the pandemic year.
Have they considered the confounding factors in the design and analysis?
The retrospective nature of this observational cohort study is subject to the flaws of all retrospective studies. The inability to test for SARS-CoV-2 antibodies may have led to an underestimation of the impact of the virus on the reported outcomes.
Was the follow-up of subjects complete enough?
Not applicable given the retrospective and observational nature of this study.
Because of the retrospective nature of the study, if data had not been recorded in the ED notes, it was not available to researchers.
It is also possible that some cases may have been missed if patients moved. However, because of the strictly defined catchment areas for each centre, we’d expect most, if not all, presentations with DKA to be recorded.
Movement restrictions were imposed during the pandemic phase (and, therefore, during the study period). The number of missed presentations because of population movement was probably low.
Was the follow-up of subjects long enough?
The study included two well-defined periods, one during the first wave of the COVID-19 pandemic and one in the preceding year.
Do you believe the results?
Yes – the study included a large patient cohort from several geographical areas. Site leads looked at retrospective data using chart review with clearly defined criteria to minimise bias.
This study also concurs with other published data from similar studies.
Can the results be applied to the local population?
Yes. The study included many sites across the PERUKI network. The results will likely translate to the wider patient population with similar patient cohorts.
Do the results of this study fit with other available evidence?
NPDA data reported a 20.7% increase in new-onset diabetes in the first COVID pandemic year, corresponding closely to DIMPLES.
This study had a different outcome from the REPEM EPISODES study – there was a true increase in new-onset diabetes and incidence and severity of DKA across the PERUKI network.
The increase in PICU admissions matches a similar increase seen by paediatric intensivists in the UK and Éire PICANET database. Birkebaek et al., in the Lancet Diabetes and Endocrinology 2022, stated that an ‘additional pronounced increase during the COVID-19 pandemic were general phenomena observed in most registries from different countries and continents.’
What are the implications of this study in practice?
This study predicted the likelihood of further years of increased incidence of new-onset diabetes. This is especially likely if SARS-CoV-2 has triggered the process of seroconversion in susceptible children. The NPDA audit suggests that 2021-2022 was another high incidence year, higher even than the first COVID pandemic year.
DKA rates are exceptionally high, despite awareness campaigns. This may be due to an increased awareness of the symptoms of new-onset diabetes coupled with decompensation of pre-existing diabetes. Healthcare workers who come in contact with children regularly will be better equipped to recognise and advocate for those who may be at risk.
Because of the multifactorial nature of the disease, it has always been difficult to establish a connection between viruses and the pathogenesis of diabetes or its severity at presentation. This evidence, related to SARS-CoV-2, may help to show that infectious triggers may trigger the onset of the disease in genetically predisposed children or accelerate the diagnostic process. Unfortunately, the lack of antibody testing in the paediatric population makes a connection harder to establish.
What can we take away from the DIMPLES study?
COVID-19 seems to have been an accelerator or precipitator for the onset of diabetes and the incidence and severity of DKA across the paediatric population. Prospective studies are required to understand further the true role of SARS-CoV-2 in the pathogenesis of diabetes (and other diseases).
A note from the lead author: Dr Caroline Ponmani
It was simply amazing to be able to do a speciality study in the Children’s Emergency Department during peak COVID. As it turned out, DIMPLES was the largest UK paediatric pandemic diabetes study. DIMPLES tells the story of paediatric diabetes in the COVID pandemic, the unusually high incidence of new-onset diabetes, the severity of DKA, the difficulties in proving causation in a complex chronic condition, and the need to recognise association and raise awareness.
We have a ringside view in the Children’s Emergency Department – the unique opportunity to see different types of acute presentations, whether DKA, sepsis or asthma, amongst a host of febrile children. DIMPLES came about due to simply observing the unusual presentations of new-onset diabetes and the severity of DKA on the shop floor in the first pandemic wave.
Our patient stories and data are powerful, and together we can make a difference as we did through DIMPLES, which really was paediatric diabetes as seen through the eyes of the children’s Emergency Medicine clinician. We want more awareness, more kindness, and more investment for the children who have this lifelong condition, and I can tell you that we are already getting there!
Declaration – Michael Barrett was a co-author of the DIMPLES study.
Birkebaek NH, Kamrath C, Grimsmann JM, Aakesson K, Cherubini V, Dovc K, de Beaufort C, Alonso GT, Gregory JW, White M, Skrivarhaug T. Impact of the COVID-19 pandemic on long-term trends in the prevalence of diabetic ketoacidosis at diagnosis of paediatric type 1 diabetes: an international multicentre study based on data from 13 national diabetes registries. The Lancet Diabetes & Endocrinology. 2022 Nov 1;10(11):786-94.
Bressan, S., Buonsenso, D., Farrugia, R., Parri, N., Oostenbrink, R., Titomanlio, L., Roland, D., Nijman, R. G., Maconochie, I., Da Dalt, L., Mintegi, S., & Country Leads (2020). Preparedness and Response to Pediatric COVID-19 in European Emergency Departments: A Survey of the REPEM and PERUKI Networks. Annals of emergency medicine, 76(6), 788–800. https://doi.org/10.1016/j.annemergmed.2020.05.018
Kanthimathinathan, H.K., Buckley, H., Davis, P.J. et al. In the eye of the storm: impact of COVID-19 pandemic on admission patterns to paediatric intensive care units in the UK and Eire. Crit Care 25, 399 (2021). https://doi.org/10.1186/s13054-021-03779-z
Nijman, R. G., Honeyford, K., Farrugia, R., Rose, K., Bognar, Z., Buonsenso, D., Da Dalt, L., De, T., Maconochie, I. K., Parri, N., Roland, D., Alfven, T., Aupiais, C., Barrett, M., Basmaci, R., Borensztajn, D., Castanhinha, S., Vasilico, C., Durnin, S., Fitzpatrick, P., … in association with the REPEM network (Research in European Pediatric Emergency Medicine) as part of the EPISODES study group (2022). Presentations of children to emergency departments across Europe and the COVID-19 pandemic: A multinational observational study. PLoS medicine, 19(8), e1003974. https://doi.org/10.1371/journal.pmed.1003974
Ponmani, C., Nijman, R. G., Roland, D., Barrett, M., Hulse, T., Whittle, V., Lyttle, M. D., & Paediatric Emergency Research United Kingdom and Ireland (PERUKI) (2023). Children presenting with diabetes and diabetic ketoacidosis to Emergency Departments during the COVID-19 pandemic in the UK and Ireland: an international retrospective observational study. Archives of disease in childhood, archdischild-2022-325280. Advance online publication. https://doi.org/10.1136/archdischild-2022-325280
Tucker, M.E. (2022). Increase in New Onset Type 1 Diabetes, DKA In Kids During Covid Pandemic. Medscape Review Article. August 18 2022. Available at https://www.medscape.com/viewarticle/979381 (Accessed: May 22 2023).