Patch
This is part of the draft I wrote lost night. I've revised it but put my conclusion first.
I'm very afraid that some people will interpret this thread as meaning that being put onto insulin is the last nail ( many people already think this )and frighten them away from a necessary medication that may lengthen their lives and reduce the risks of complications. Individuals have great success on insulin and this includes many T2s (have a look at some of the US forums , some have regular T2 posters who have been on insulin for 20+ years)
This was not a controlled trial, it cannot establish any causality merely suggest hypotheses.
If we took a group of newly diagnosed T2diabetics and randomly put one group onto insulin and the other on metformin for 3 years (the time looked at in this analysis ) would the insulin users develop more complications or die more frequently than those using metformin ?
I suggest that there would be a difference in glucose levels and that would depend on the nature of the subjects diabetes. (after all the medications work in different ways and address different problems. I suspect that individual results would depend upon how much glucose was produced from the liver.(met treats this, insulin doesn't, that's a reason they work together)
A medscape reviewer suggests that there are three possible reasons for this pattern of results: http://www.medscape.com/viewarticle/804103
- Residual confounding
- that the co-morbidities are interrelated (well I certainly think they are)
- There is one precipitating event. The researchers consider this to be the case ie insulin use .
The researchers have tried to eliminate confounders but include the caveat
these results must be considered with the caveat of residual confounding both within the candidate covariates and those factors that could not be represented within our models
Here's just a couple of caveats I have .
They haven't taken into account the prior effects of longer term higher glucose levels
They acknowledge progression and the late use of insulin in many cases they don't seem to control for this aspect ( I can't see how they can)
Metabolic 'memory' may be important, just as prior good control seems to have a lasting beneficial effect when control deteriorates the opposite is also true.
T1s who omit insulin and run intermittent high glucose levels often don't see problems until later. These frequently occur when they have developed better control. It isn't the insulin that causes the problems but the damage done earlier. (early worsening of retinopathy is slightly different example)
When considering an outcome, they excluded people who already had markers for this problem ; so for kidney disease they removed those with known poor results on kidney tests
but they could only do this accurately if there was a test on record, we don't actually know the state of the kidneys in those with no test. ie the missing data problem ( This is GP data and we also know from the diabetes audit that at least 30% of people with D in the UK don't receive this test. In some areas and GP practices it is considerably more)
Excluding those with known large vessel disease from CVD outcomes doesn't account for the interelationship between various microvascular and macrovascular complications ,
those who have kidney disease may well die from a heart attack or stroke.
The only long time trial of insulin use is the DCCT (so T1) This did look at insulin doses and future complications. Those who took higher doses had similar problems to those often associated with T2 . They had higher BMI and the various markers of the metabolic syndrome.
Weight, symptoms of the metabolic syndrome or
subsequent higher insulin doses were not predictive of future complications, however lowered kidney function was predictive of both micro and macrovascular complications. (back to that interrelationship again)
http://www.ncbi.nlm.nih.gov/pubmed/17327345
Finally, Prof Currie, the lead researcher also said
"
The vast majority of people who take insulin will experience no adverse effects and it remains a reliable and common form of treatment worldwide, but this study shows that we need to investigate this matter urgently and the drug regulatory authorities should take interest in this issue"
http://www.bbc.co.uk/news/uk-wales-21319226
and that is exactly what needs to be done ie further research because this study can only delve into possibilities and not establish anything.