Thread by @dgurdasani1

Lots of people dismissing links between COVID-19 and all-cause diabetes. An association that's been shown in multiple studies- whether this increase is due to more diabetes or SARS2 precipitating diabetic keto-acidosis allowing these to be diagnosed is not known. A brief look👇

Peer-reviewed study from Imperial researchers published in Diabetes Care. This study showed an 80% increase in presentation of new cases of type I diabetes under 17 years between March 23rd and June 4th 2020.

pubmed.ncbi.nlm.nih.gov/32816997/

Most presented with an acute form of diabetes called diabetic keto-acidosis (DKA). DKA is an acute presentation of diabetes- which can be life-threatening. It's usually caused or precipitated by certain factors, including infections.

So there are 2 possibilities:
1) they were presenting with type 1 diabetes more and being diagnosed more because COVID-19 was precipitating DKA
2) type 1 diabetes was occurring more due to direct effects from SARS-CoV-2 infection

Although this is an 80% increase, numbers were small, so still a rare disease. But a concerning association nevertheless whether COVID-19 precipitating DKA or linked to an increase in incidence of type I diabetes.

As the study outlines, DKA is a serious complication- one SARS-CoV-2 +ve child presenting with DKA had a cardiac arrest, required ventilation, and recovered subsequently.

The study concludes:
"SARS-CoV-2 exposure contributed to the observed increase in cases by precipitating or accelerating type 1 diabetes onset."

Was this a perfect study? No, not at all. It's essentially looking at real-world inpatient data from 5 units in London that are part of the NWL Pediatric Diabetes Network - and seeing if new cases of diabetes diagnosed between 23rd March and 4th June 2020 were as expected

Not even all patients who were admitted were tested for SARS-CoV-2 - PCR or antibody. Why was it published then? Because observations like this are important, and give us an idea of areas that need more research to establish if early observations are causal links or not.

Several other studies - in adults and children have shown concerning patterns. Some show increase in new diagnosis of diabetes, or increased severity of diabetes at presentation or longer duration of disease. As well as links with increased mortality among those newly diagnosed.

Here's a comprehensive review showing evidence - in children and adults. The pathogenesis is unclear, but it's a consistent association observed - with either incidence, or severity of newly diagnosed diabetes.
ncbi.nlm.nih.gov/pmc/articles/P…

As the study outlines, there isn't clarity whether this is type I diabetes, type II diabetes or a new form precipitated by COVID-19 or existing diabetes being diagnosed due to increased severity with COVID-19. This area needs urgent study, for us to be able to understand more.

Here's a study of 5.8 million men from the US veteran's health programme showing that newly diagnosed diabetes (presumably all types) increased in those post-COVID-19 by 8x compared to those not known to be infected.

The study for each disease examined, excluded those who had been diagnosed with it before, so it could look at newly diagnosed cases. There was a gradient with severity of acute illness. While those being hosp being at high risk even those with 'mild' illness were at greater risk

The authors did some analyses to assess the bias/confounding within their studies, including looking at illnesses that it wouldn't make sense to be associated with COVID-19 and found no association.

Although I suspect the VHA programme regularly examines patients, I'm not sure we can entirely rule out that perhaps some of those in the COVID-19 group had undiagnosed diabetes which was diagnosed post-COVID-19 (would have to be to a greater extent than the uninfected gp)

But in the absence of randomised trials, which of course can't be done with SARS-CoV-2, this is the best evidence we have, and we shouldn't ignore it. Rather this should be impetus for further study.

Worth noting that there are several case reports now of new diagnoses of diabetes or DKA following infection with SARS-CoV-2. We still don't understand what contributes to this, as this piece in the NEJM points out:
nejm.org/doi/full/10.10…

And of course the much critiqued CDC study that came out recently suggesting higher risk of all-cause diabetes in children.
cdc.gov/mmwr/volumes/7…

Are these studies perfect? No. This is what a lot of real-world data looks like. But even the most carefully conducted studies with controls and adjustment for confounders (e.g. the VHA study) show a consistent pattern. Does this mean the association is causal? We don't know.

Should we ignore them? No
Should the precautionary principle apply as we study it more? Yes, absolutely.

When many observational studies point the same way -, we do need to practice caution mass exposing people to a novel threat, as we understand it's impact better.

It's worth remembering that even if some of the associations for incident diseases we're seeing with COVID-19 are causal, this would mean massive increase in chronic illness at population level in the next few years. Is this a risk we're willing to take?