I'm also concerned at suggesting that it is necessary to have an HbA1c in the 4s to be truly normal, whatever Dr Bernstein says there is no objective evidence to say this is the case .
Acutally, very few people have HbA1c levels of below 4.5 % and these are often associated with health problems. ( some types of anaemia can cause low HbA1cs)
David Mendosa interviewed the head of the NGSP (the organisation that standardised tests) He said that that in non diabetics 99% of the values are between 4.5 and 6% .95% are between 4.7% to 5.7% with a slight skew towards the higher end
http://www.healthcentral.com/diabetes/c/17/59130/normal-a1c-level/
see also Low Hemoglobin A1c and Risk of All-Cause Mortality Among US Adults Without Diabetes where the least risk was at an HbA1c of between 5% and 5,4% (on either side it increased , levels below 4.0% were actually a far bigger risk factor than those of 6-6.4%)
http://circoutcomes.ahajournals.org/content/3/6/661/T2.expansion.html
It also isn't normal to stay at below 5mmol/l all day long .
The non diabetics in the International trial from which the estimated average glucose levels are derived wore continuous monitors for part of the time. 100% of them went above 6.1%, 99% of them went above 7mmol/l and 93% of them above 7.8mmol/l.
Whilst the median time above 7.8mmol/l was only 31mins, there were people within the middle group (interquartile range) that spent up to 81 min at this level.
The median time that all of the people in the trial spent above 6.1% was almost 7 hours (395 mins) , sonot a short time (IQR 273-688min)
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2892065/?report=reader#!po=50.0000
Lastly, estimated average levels given for any HbA1c are just that:
estimated and average. One persons 5% may reflect very different real average glucose levels than anothers.You'll find a convertor on this site using the algorithm worked out from the trial mentioned above. In the UK and many countries you won't get given that estimated average along with your HbA1c result.. That's because for many people it won't be right
At each end of the spectrum there are people who glycate more or less haemoglobin for any given glucose level. .
Individual variation means that someone with a true average glucose level of 9.4mmol/l could have an HbA1c as low as 6.5% or as high as 9%
http://care.diabetesjournals.org/content/32/1/e11.long
This means that a person who is a high glycator would have to have achieve far lower average levels to achieve an HbA1c in the 5s (let alone the 4s) than someone who was lucky enough to be a low glycator. This is nothing people can do about this (may well be genetic)
Something to bare in mind when looking at different peoples HbA1c levels
Edited to add. I reckon I'm lucky enough to be a low glycator and I have had an HbA1c in the 4s and yes I was warned by professionals and people on forums that I could lose hypo awareness. When I got off my bike, felt a bit hypo and tested to find that my glucose meter said 'lo' ie too low to measure for the second time that week I realised they were right. I now keep my levels at around 5.8-6% and have good hypo awareness.
.