Hi I'm not an expert but I did a bit of looking up. This is what I found.
The fasting levels your OH has are under the cut off levels for impaired fasting glucose as defined by the World Health Organisation (6.1mmol). M.s Atkiins and Bernstein practice in the US where the ADA uses a lower level of 5.6mmol.
Because of this difference the WHO revisited it's advice in 2007 this is what they recommended
The fasting plasma glucose cut-point for Impaired Fasting Glucose (IFG)
should remain at 6.1mmol/l.
This decision was based on concerns about the significant increase in IFG
prevalence which would occur with lowering the cut-point and the impact
on individuals and health systems. There is a lack of evidence of any
benefit in terms of reducing adverse outcomes or progression to diabetes
and people identified by a lower cut-point eg 5.6mmol/l (100mg/dl) have
a more favourable cardiovascular risk profile and only half the risk of
developing diabetes compared with those above the current WHO cutpoint.
Lowering the cut-point would increase the proportion of people
with IGT who also have IFG but decreases the proportion of people with
IFG who also have IGT.
Consideration should be given to replacing this category of intermediate
hyperglycaemia by an overall risk assessment for diabetes, cardiovascular
disease, or both, which includes a measure of glucose as a continuous
variable.
The low level following the GTT shows that the insulin was at present sufficient and working well, perhaps rather too well. I did some looking up to find out what this very low level might signify and found that low levels can occasionally indicate other problems, so if there are other symptoms they should be investigated. However, I found a suggestion that the low carb diet may be responsible.
I'm not sure about how valid this particular source is (it's from the type of website I normally avoid)
but I found this
1) the gtt test used to be used to diagnose 'reactive' hypoglycaemia it the level fell below 50mg/dl (2,7mmol)
2) It wasn't a good test one reason was 'it was frequently performed improperly. 'Consider that unless a patient undergoing the procedure ate extra carbohydrates for three days prior to the test, blood sugar readings could come in abnormally low; a hefty dose of sugar following several days of a relatively low-carbohydrate diet can throw off the system. That caveat was often overlooked and led to over-diagnosis'....
http://www.encognitive.com/node/5445
further searching did show that people are told to eat a normal or often in the case of pregnant women a high carb diet before testing.
The one thing about his rising level of fasting glucose is that he is still at a level where lifestyle may prevent diabetes (exercise, weight loss, healthy diet.. I'll not define that!)