Is Covid-19 causing diabetes in kids?

A new study shed light on a complicated evidence landscape

A few months into the Covid-19 pandemic, murmurings began to emerge among the paediatric community of a noticeable uptick in the number of children presenting to hospital with a condition called “Diabetic Ketoacidosis” (DKA).

DKA is often the first presentation of diabetes in children. It occurs because due to the deficiency of insulin, the body cannot utilise glucose. The body therefore thinks it is starving, and so makes a substrate called ketones as a new form of energy. Unfortunately the body also cannot use ketones due to the deficiency of insulin. Ketones make the blood more acidic, which makes kids feel sick and vomit. The high levels of glucose in the blood cause the body to lose more water (in an attempt to discard all the excess glucose) and so the children also become more dehydrated. The condition can easily be mistaken for a severe vomiting bug, and children often present late to hospital in near critical illness.

It was initially unclear whether there was a true uptick in these numbers, or whether we just noticed them more because so few children were coming to hospital with minor ailments during lockdowns. Or, whether children with new diabetes were getting diagnosed later than usual due to lack of access to healthcare, and so a larger proportion were going into DKA before diagnosis.

Now there is data from several countries which looks convincingly as though both total numbers of children with DKA, and importantly, new diagnoses of type 1 diabetes rose significantly during the pandemic.

But why?

It’s Covid, stupid

As with many things which have gone awry over the past few years, the obvious place to point the finger is at Covid-19. Viruses do weird things, and it is easy to hand-wave at the new kid on the block and blame them for everything. Sometimes this is true, such as with the new hyperinflammatory syndrome (known as MIS-C or PIMS-TS). However, sometimes this is untrue, such as the outbreak of fulminant hepatitis in children in the spring of 2022. This turned out to be due to a combination of a particular genetic susceptibility and a virus called AAV2.

This is a great example of why we must carefully evaluate the evidence before jumping to conclusions.

Covid-19 itself could be a plausible explanation. We are still not entirely sure why some children develop type 1 diabetes, but as with many conditions, a viral trigger has certainly been proposed as a possible cause. It is not unreasonable to suggest then that Covid-19 could have a particular propensity for triggering this condition in children. This would be a worry given the virus is here to stay and is otherwise relatively harmless to most children.

Some evidence has suggested that there could be a link. A couple of studies from the US using health insurance data suggested that children who had been diagnosed with Covid-19 were more likely to be subsequently diagnosed with diabetes in the following few months (here and here).

Surely that is case closed?

Not all as it seems

There are well recognised difficulties with using US insurance based databases for observational research. The major potential issues with the two aforementioned studies in particular is reverse causation: not that Covid-19 causes type 1 diabetes, but having type 1 diabetes (undiagnosed) makes you more likely to test positive for Covid-19.

We know that the overwhelming majority of cases of Covid-19 in children went undetected during the pandemic (depending on the age of the child and time period, somewhere between 5 and 20% of children with Covid-19 actually got tested). We know type 1 diabetes is a risk factor for more severe illness and symptoms. Having undiagnosed type 1 diabetes will almost certainly make you more likely to have worse symptoms and to get tested. A study from Scotland certainly suggested that this might be the cause of a link between positive tests and diagnosis of type 1 diabetes. We cannot be certain this is what happened in the studies from the US, but we must bear this possible explanation in mind, especially in view of other evidence.

Other countries have failed to find the same association between testing positive for Covid-19 and the development of type 1 diabetes. A nationwide observational study from Denmark did not observe an increase in rates of type 1 diabetes in children following a positive Covid-19 test compared to those with only negative tests. In a huge study from Finland, only a tiny fraction of children who were newly diagnosed with type 1 diabetes in Finland during the pandemic had serological evidence of Covid-19 infection (<1%), despite a large increase in number of children being diagnosed. In a multinational cross-sectional cohort from the US and Germany looking at the prevalence of the antibodies which cause type 1 diabetes in children, there was no difference in those with or without evidence of prior Covid-19.

There is other circumstantial, epidemiological evidence. Countries in northern Europe which initially had very low rates of Covid-19 still saw big rises in rates of type 1 diabetes in children. Most studies found increases in rates of type 1 diabetes following the first year or so of the pandemic; but actual rates of infection of children in this time period were generally quite low; between 1 - 10% depending on the country. There were periods in the UK during the omicron wave in January 2022 where 10% of all children were testing positive at the same time. We would have expected a tidal wave of new cases of diabetes following this wave, which as of yet no evidence has emerged. Note was made in a study from Germany that the rates of new cases of type 1 diabetes did not seem to match the rates of Covid-19 infections in different waves.

All of this would suggest Covid-19 as unlikely to be the cause, but what we really need is a prospective study which regularly tests children for both antibodies to SARS-CoV-2, and evidence of type 1 diabetes. As luck would have it, that’s exactly what we now have.

The new study

A new, multinational prospective cohort study has been published, following >4,500 children aged 9 to 15 years from January 2020 to December 2021 in the US, Sweden, Finland and Germany. They were tested at regular intervals for antibodies which can cause type 1 diabetes (islet autoantibodies), tested for type 1 diabetes itself, and for antibodies against SARS-CoV-2 during this period.

During this time, around 15% of children developed antibodies against SARS-CoV-2 infection. Of the children without islet autoantibodies to begin with, 1% of them (40 children) developed these antibodies during the study. Only 5 of these children ever had evidence of having Covid-19, all of which after they had already developed islet autoantibodies.

Of the 4,500 children, 45 were diagnosed with type 1 diabetes during the study period. There was no difference in rates of diagnosis of type 1 diabetes between the children with or without a history of SARS-CoV-2 infection.

This is by far the highest quality evidence we have on this topic to date, however there are some limitations. The population is relatively small given that developing type 1 diabetes is quite rare. The age range is quite narrow at 9 to 15 years, and the children chosen for this cohort study are known to be genetically at higher risk of developing type 1 diabetes. Never-the-less, these findings are extremely reassuring.

So what is the cause?

The truth is that it is still unclear what the cause is (or was). We cannot rule out Covid-19, although it is looking less and less likely. Groups in Germany and Finland have suggested it is a result of indirect pandemic effects - environmental factors triggered by lockdowns/reduced socialisation and/or reduced circulation of other normal microbes which impact childhood immunity. Additional follow up and epidemiological investigation will hopefully bring us closer to understanding, but there is always the possibility that this was a transient phenomenon which will remain a mystery for the foreseeable future.

Summary

There was a significant rise in the number of children being diagnosed with type 1 diabetes during the first years of the pandemic. Some observational evidence has suggested the risk was increased following SARS-CoV-2 infection, but this is contradicted by other large scale observational studies, serological evidence, circumstantial epidemiological evidence, and now a new, high quality prospective cohort study. The true cause remains unclear, but may be secondary effects of the restrictions put in place during the pandemic to reduce circulation of Covid-19.