Article: Diabetes sugar 'can go too low'

The_Conductor

http://news.bbc.co.uk/2/hi/health/8481770.stm

Intense treatment to lower blood sugar in patients with diabetes could prove nearly as harmful as allowing glucose levels to remain high, a study says.

Cardiff researchers looked at nearly 50,000 patients with type 2 diabetes and found the lowest glucose levels linked to a heightened risk of death.

Significant differences in death rates between patients on insulin and those taking tablets are also flagged up.

But there could be various explanations for this, experts noted.

Patients taking insulin-based treatments have been urged not to stop taking their medication as a result of the Cardiff University study, which is published in The Lancet.

Changing treatments

Using data from GPs, the team identified 27,965 patients aged 50 and above with type 2 diabetes whose treatment had been intensified to include two oral blood glucose lowering agents - metformin and sulphonylurea.

A further 20,005 patients who had been moved on to treatment which included insulin were added to the study.

Patients whose HbA1c levels - the proportion of red blood cells with glucose attached to them - were around 7.5%, ran the lowest risk of dying from any cause.

For both groups this risk went up by more than half if levels dropped to 6.4%, the lowest levels recorded. For those with the highest levels the risk of death increased by nearly 80%.

But the risks appeared to be particularly pronounced among those on the insulin-based regimen than those on the combined treatment.

Irrespective of whether their HbA1c levels were low or high, there were 2,834 deaths in the insulin-taking group between 1986 and 2008, nearly 50% more than in the combined group.

'Don't stop'

The authors acknowledged there could be various factors associated with this, such as these being older patients with more health problems, who perhaps had had diabetes for a longer period of time. They also make reference to a possible link between use of insulin and cancer progression that had been reported in a different study.

"Whether intensification of glucose control with insulin therapy alone further heightens risk of death in patients with diabetes needs further investigation and assessment of the overall risk balance," wrote lead author Dr Craig Currie.

"Low and high mean HbA1c values were associated with increased all-cause mortality and cardiac events. If confirmed, diabetes guidelines might need revision to include a minimum HbA1c value."

These findings were in line with those of a major ongoing trial in the US, which pulled patients off a regimen of intensive blood sugar management after noting an unexpected increase in total deaths among this arm of its study.

Dr Iain Frame, head of research at Diabetes UK, described this latest study as "potentially important" but stressed it had limitations.

"It is not clear what the causes of death were from the results reported. Furthermore, when it comes to the suggestion made in this research that insulin could increase the risk of death, we must consider important factors such as age, the duration of their diabetes and how the participants managed their condition."

While people would be able to manage their condition for a period with diet, exercise and even tablets, many would eventually have to move on to insulin, he noted.

"We would advise people with type 2 diabetes who use insulin not to stop taking their medication. However, if they are worried about blood glucose targets, they should discuss this with their healthcare team."

mike65

Patients whose HbA1c levels - the proportion of red blood cells with glucose attached to them - were around 7.5%, ran the lowest risk of dying from any cause.

For both groups this risk went up by more than half if levels dropped to 6.4%, the lowest levels recorded. For those with the highest levels the risk of death increased by nearly 80%

Bloody hell my endo won't like this!

lucernarian

The study which highlighted a statistically significant correlation between cancer development and insulin, that was only about Lantus and not insulins in general, right?

I'd like to see the outcome if mortality rates for diabetics taking insulin and who were below the age of 60 were studied, for example.

This certainly needs more research, and specifically into ruling in or out any other possible correlations.

The_Conductor

If you look at the study notes- there were more people on oral blood lowering glucose treatments, than were on insulin itself. As such- I'm not sure that the study was specifically studying insulin- more so the HbA1Cs (regardless of how people arrived at those scores).

Of more interest- particularly given the age demographics of the group- is what the state of health of the individuals concerned was, prior to commencement of a particular course of treatment (be it oral tablets or injectable insulin). Normally Lantis would only be prescribed to the worse patients- possibly those who have had diabetes for a longer period of time- or are viewed as needing assistance in controlling it (possibly because of other health factors).

You really need to have a reasonable picture of an average member of the Lantis group- versus an average member of the oral medication group- at the start of the trial- if you want to try to derive an indication of the likliness of a particular course of intervention to be detrimental to that person's health. Also pertinent- would be initial BMIs, initial HbA1Cs and some sort of an idea of the level of activity and diet of those concerned.

While the article is of concern- the way in which its reported raises more questions than it answers, and risks sensationalising a study- that may have altogether more mundane findings than the journalist is suggesting.

lucernarian

Many Type 1s however would use Lantus on a regular basis, but their age/health profile would often be different, so perhaps the different cases from the Lantus study need to be separated out?

BC

I'm type 1 on Lantus. My consultant says he has studied the results from these and reckons there are just way too many factors involved and way too many other things that could have happened to get worried about them.

Carsinian Thau

BC wrote: »

I'm type 1 on Lantus. My consultant says he has studied the results from these and reckons there are just way too many factors involved and way too many other things that could have happened to get worried about them.

I'm open to correction on this, but I'd assume that the study wouldn't even apply to you as a Type 1.

It only looked at Type 2 diabetics and our condition is rather different from theirs.

The_Conductor

Carsinian Thau wrote: »

I'm open to correction on this, but I'd assume that the study wouldn't even apply to you as a Type 1.

It only looked at Type 2 diabetics and our condition is rather different from theirs.

Correct- and further it was solely for those over the age of 50- another reason the study isn't applicable to either of you......

In my opinion it is high time that there was far more differentiation made between Type 1 and Type 2 diabetes (and please don't get me started on the youngsters with type 1.5.........)

colrow

Well its got me a bit worried, as I've been concentrating on keeping my bg in the low 6's 6.2 - 6.4 and at one time i was hoping to get it in the 5's, I'm seeing my doc on Wednesday and I'll take a copy to him and see what his opinion is, I'll let you know what he says.

The_Conductor

colrow wrote: »

Well its got me a bit worried, as I've been concentrating on keeping my bg in the low 6's 6.2 - 6.4 and at one time i was hoping to get it in the 5's, I'm seeing my doc on Wednesday and I'll take a copy to him and see what his opinion is, I'll let you know what he says.

Seriously- the brief writeup on the study above- is insufficient in nature to draw any conclusions whatsoever.

It involved Type 2s, over the age of 50- and there is no details of their previous health prior to the study. Further- there is no indication whatsoever as to whether there was an age related profile or a BMI related profile associated with the different treatments prescribed or the differing mortality rates reported.

In the age group in particular- heart disease is almost 4 times as likely to be a stated COD, than diabetes related complications- this is something you should keep in mind.

Do not automatically assume that any study you hear of on the internet is directly relevant to you- 9-times-out-of-10, it most probably is not.........

lucernarian

Clarification needed here:

The Lantus study which detailed a possible link with cancer in a cohort study is different to the one above. The Lantus study was open to both type 1 and 2 patients. As the number of type 1 patients exclusively on Lantus would be fairly low, I would expect a large majority of the Lantus-only users were type 2.

The above study dealt with exclusively Type 2 patients over 50 years of age.