Insulin inhalers don't prevent Type 1 diabetes in children with antibodies

Nasal insulin to prevent type 1 diabetes in children with HLA genotypes and autoantibodies conferring increased risk of disease: a double-blind, randomised controlled trial
Original source: The Lancet, Volume 372, Issue 9651
Authors: Kirsti Näntö-Salonen (corresponding author), Antti Kupila, Satu Simell, Heli Siljander, Tiina Salonsaari, Anne Hekkala, Sari Korhonen, Risto Erkkola, Jukka I Sipilä, Lotta Haavisto, Marja Siltala, Juhani Tuominen, Jari Hakalax, Heikki Hyöty, Jorma Ilonen, Riitta Veijola, Tuula Simell, Mikael Knip, Olli Simell
Publication date: 15 November 2008
DOI (digital document ID): 10.1016/S0140-6736(08)61309-4
Requires subscription? Yes
Publicly-available abstract: Yes

Synopsis:
In mice engineered to have risk of Type 1 diabetes, inhaled insulin has been shown to reduce their likelihood of developing diabetes. The authors wanted to know if this would work in humans. They screened a number of infants for a specific gene (the HLA-DQB1 susceptibility allele) linked to increased risk of developing Type 1 diabetes. The siblings of susceptible infants were also screened. Of these, over 11,000 infants and 1500 related siblings consented to having their children's blood screened for diabetes at 3-12 month intervals. [One would presume that for minors, their parents would have been the ones giving consent.] From these, 224 infants and 40 of their siblings (all of whom tested positive for two or more diabetes-specific antibodies) were entered into a trial in which some would be given short-acting human insulin (1 unit/kg) or placebo once a day intranasally (inhaled) until they developed Type 1 diabetes and needed traditional injectable insulin. The children were studied for an average length of 1.8 years (range lasted from 0-9.7 years). The study ended early because the insulin did not show any benefit -- as many children on insulin as on placebo developed diabetes, over similar ranges of time from the start of the trial.

Reliability:
Almost all of the authors are either medical doctors or professors of medicine. The number of children studied was not very large (only about 260), and they had already developed diabetes-related antibodies.

Would the results have been different if insulin had been administered before antibodies had developed, with the trial endpoint being the development of antibodies? Would the results have been different if a basal insulin had been used, or a different dosage?

As far as I know there is no inhaled insulin treatment for Type 1 diabetes in people. Does inhaled insulin work as primary treatment in those of that strain of mice who have developed diabetes? If it does, then the experiment is "comparing apples and oranges", and the analogous human experiment would require injected insulin.

Practical Applications:
Had there been a reduction in the number of children studied who developed Type 1 diabetes, or a significant delay in their development of Type 1 diabetes, then inhaled insulin therapy could have been used as a preventative (prophylactic) measure. The negative findings show that it cannot be.

Ethics Note:
The study was registered with ClinicalTrials.gov, number NCT00223613. Since the subjects were likely minors, their parents would have been required to provide the necessary consent.

This synopsis was based on the abstract only.

Can genetic risks accurately predict Type 2 diabetes?

Clinical Risk Factors, DNA Variants, and the Development of Type 2 Diabetes
Original source: New England Journal of Medicine Volume 359, Number 21
Authors: Valeriya Lyssenko, Anna Jonsson, Peter Almgren, Nicoló Pulizzi, Bo Isomaa, Tiinamaija Tuomi, Göran Berglund, David Altshuler, Peter Nilsson, and Leif Groop
Publication date: November 20, 2008
DOI (digital document ID): not available
Requires subscription? Yes
Publicly-available abstract: Yes

Synopsis:
The aim of this paper is to determine whether environmental and clinical risk indicators, genetic factors, or both can be used to predict whether or not a person will develop Type 2 diabetes.

The authors singled out 16 specific gene markers that might be used to predict risk or development of diabetes. They tracked those, plus known environmental and clinical risk indicators, for over 16,000 Swedish patients and over 2500 Finnish patients, for a median period of 23.5 years. They studied the effect of variations of the gene markers on insulin secretion and insulin action over this period of time.

Over the observation period, approximately 11.7% of the subjects developed diabetes. Variations in 11 of the genes studied were associated with developing diabetes, independent of other risk indicators. Variants in eight of those genes were specifically associated with impaired beta cell function.

Adding the genetic risk to the clinical and environmental indicators slightly improved the long-term ability to predict whether or not an individual would develop Type 2 diabetes. This predictive ability increased with the duration of the follow-up.

Reliability:
While a large population was observed, and the medium observation time was over 20 years, the abstract does not provide insight into the distribution of follow-up duration, to let us know whether the median is a statistically-significant number. The improved prediction ability turns out to be about 1.5% better than the accuracy without the genetic information, which may not be cost-effective at the present time.

Practical Applications:
Known genetic risk is sometimes an incentive for making lifestyle changes to reduce the other risks for developing diabetes.

If the full paper shows direct correlations between insulin secretion, insulin resistance, and genetic variations, then initial pharmaceutical therapy can be better targeted to the individual's needs.

This synopsis was based on the abstract only.

A possible genetic component of Type 2 diabetes

Variance of the SGK1 Gene Is Associated with Insulin Secretion in Different European Populations: Results from the TUEF, EUGENE2, and METSIM Studies
Original source: PLoS ONE 3(11): e3506.
Authors: Björn Friedrich, Peter Weyrich, Alena Stančáková, Jianjung Wang, Johanna Kuusisto, Markku Laakso, Giorgio Sesti, Elena Succurro, Ulf Smith, Torben Hansen, Oluf Pedersen, Fausto Machicao, Silke Schäfer, Florian Lang, Teut Risler, Susanne Ullrich, Norbert Stefan, Andreas Fritsche (corresponding author), Hans-Ulrich Häring
Publication date: 5 November 2008
DOI (digital document ID): 10.1371/journal.pone.0003506
Requires subscription? No
Publicly-available abstract: Yes

Synopsis:
The authors used a "classical candidate-gene approach" to examine genes that play a role in insulin secretion and their potential roles in the development of insulin-resistant diabetes. The gene they focused on is serum and glucocorticoid inducible kinase SGK1, which is encoded by the gene SGK on chromosome 6q23. The action of this gene is increased in the presence of corticoids and insulin, among other things, and seems to provide a molecular link between electrolyte activation and glucose stability. It also appears to play a role in both insulin signalling and insulin secretion and blood pressure.

The investigators looked at genetic variance in SGK, and potential associations with insulin secretory function, in two specific groups of people (the 725-person TUEF cohort and the 634-person EUGENE2 participants) whose diabetes risks had been examined for insulin secretion traits among those who were prediabetic. These were compared with the results of a study on an older, genetically-similar population (the 4457-person METSIM cohort) to determine whether their hypothesis was correct.

Insulin sensitivity and insulin production were measured using oral glucose tolerance tests (OGTT), measuring c-peptide levels or plasma insulin levels to determine insulin secretion, along with blood glucose levels to determine insulin sensitivity. BMI and waist-hip ratio were also measured in these populations. The patients were correlated to specific variations of one of four tagged alleles (rs9402571) of the SGK-1 gene.

There was no association of the specific tagged alleles with insulin sensitivity in either the TUEF screening population or in the EUGENE2 study, but the authors suggest that this expression of the gene's expression is based on BMI. They mention that the spot in which the SGK1 gene sits has been "associated with BMI progression in participants of the Framingham Heart Study."

On the other hand, the TUEF and EUGENE2 studies did report an increase in insulin secretion in lean patients with one specific variation of the rs9402571 allele. METSIM participants with this variation had a lower incidence of Type 2 diabetes. The authors suggest that this variation may protect against Type 2 diabetes, but that its effect may become blocked or blurred in overweight or obese individuals.

Their analysis suggests that the rs9402571 allele of the SGK1 gene is associated with altered insulin secretion and with Type 2 diabetes.


Reliability:
The authors are associated with accredited medical institutions and the results were published in a peer-reviewed journal. The subsets of the TUEF and EUGENE2 participants' data analyzed in the paper are relatively small. The correlation with an older population of similar genome poses a good initial check on the original hypothesis; the ongoing nature of the TUEF study (Tübingen Family Study) suggests that a more definitive correlation will be possible at some time in the future.

Practical Applications:
The developing discipline of genomics suggests that we might one day be able to base our healthcare on how our own particular genes work.

This synopsis was based on the entire article.