Pancreatic islet cell response to hypoglycaemia in long standing type 1 diabetes

Type 1 diabetes mellitus (T1D) is an autoimmune disease in which pancreatic *-
cells are destroyed and endogenous insulin production is lost. The long term
detrimental effects of the ensuing chronic hyperglycaemia can be partially
prevented by intensive insulin therapy. However, only 15% of the patients with
T1D reaches the target HbA1c of 53 mmol/mol, with hypoglycaemia being the most
important limiting factor. The high burden of hypoglycaemia in patients with
T1D can be explained by exogenous insulin therapy combined with dysfunction of
the counter regulatory mechanisms against hypoglycaemia.

Failure of the *-cell undermines two out of the three known counter regulatory
mechanisms, namely the ability to lower the insulin secretion and increase the
glucagon release in response to hypoglycaemia. This is illustrated by the
finding that higher levels of remaining endogenous insulin secretion in the
first few years after diagnosis are associated with fewer hypoglycaemic events
in combination with a lower risk of long term complications. There is mounting
evidence that some *-cells still survive in a significant portion of patients
with long standing T1D. In patients with the diagnosis of T1D for 4-67 years, *-
cells are detected in 88% of subjects histologically and stimulated C-peptide
concentrations could be detected with an ultrasensitive assay in 73% of
patients with T1D and median disease duration of thirty years.2;3 It is not
known whether this small population of surviving *-cells can still contribute
to hypoglycaemia counter regulation by decreasing insulin secretion in reaction
to hypoglycaemia.

Continuous glucose monitoring and flash glucose monitoring can help reduce the
burden of hypoglycemia, however it is also known that these devices are less
accurate in hypoglycemia, and there is a time delay for interstitial
measurements when compared to capillary measurements. It is not known how much
this influence the utility of these devices during the emergence and resolution
of hypoglycaemia.

People living close with diabetes patients frequently mention that they can see
from the facial expression whether that person is experiencing a hypoglycaemia,
often before the patient notices him or herself. Our goal is to explore if
during hypoglycaemia with machine learning techniques a hypoglycaemia face or
*hypo-face* could be identified. This tool might help patients in the future to
recognize personal hypoglycaemic symptoms and might be used as an education
tool in gaining hypoglycaemia awareness.

Moreover a pilot study was performed in type 1 diabetes in which was found that
heart rate variability could be used in detecting occurrence of hypoglycaemia.
We would like to investigate how this method perform both in accuracy and
timing against the glucose measurement devices and the *hypo-face*.

- To determine whether the remaining *-cell in long standing type 1
diabetes are capable of decreasing their insulin secretion upon hypoglycaemia
- To determine how much delay there is between plasma glucose and interstitial
glucose measured by a Flash glucose monitor and an implantable continuous
glucose monitor during the emergence and the resolution of a hypoglycaemia
- To explore if specific changes in the face during hypoglycaemia in type 1 DM
could be recognized by machine learning techniques and thereby creating a tool
that can identify a *hypo *face*.
- To determine how the predicting value of the *hypo-face* is in comparison to
predicting hypoglycaemia with heart rate variability
- To determine whether the *hypo-face* and change in heart rate variability are
associated with the different counter regulatory mechanisms of hypoglycaemia

This is a prospective, single-centre, non-therapeutic intervention study

The participants will visit our research unit after an overnight fast, where
they will receive a fixed insulin infusion and a variable glucose infusion at
t=0 to bring the participant in an euglycemic state (5 mmol/L) (t=0-60min). For
this purpose, frequent plasma glucose concentrations will be measured. At t=105
minutes the glucose infusion is decreased to allow the glucose concentration to
drop to 2.5 mmol/L and then at t=210 min 30 grams of oral dextrose is
ingested.

Burden:
The participants are asked to visit once after an overnight fast and will
remain recumbent for approximately 5.5 hours. Two venous catheters will be
placed to infuse at average 450ml and to draw at maximum 100 mL blood. The
participants will probably experience some symptoms related to sympathetic
neural activation by hypoglycaemia including palpitation, tremor, sweating and
paresthesia.

Risks:
By inducing the hypoglycaemia in a very controlled setting with very frequent
glucose measurements, the risk of severe hypoglycaemic symptoms is very low. In
the unlikely event that a severe hypoglycaemic event does occur, intravenous
glucose can immediately be administered to counter the hypoglycaemia.

Benefits:
The study will give more insight in how novel devices perform in detecting
hypoglycaemia and looks into possible future approaches of detecting
hypoglycaemia. It will also create more knowledge in the pathophysiology of
hypoglycaemia in patients with diabetes mellitus type 1 and thereby possible
future therapeutic approaches.