Reply from Roy Taylor wrt my queries on ND

Oldvatr

Oldvatr

Expert
Messages
8,470
Type of diabetes
Type 2
Treatment type
Tablets (oral)
Recently we had several threads discussing aspects of the Newcastle Diet Study, and whether there might be any systemic error in some of the test procedures.
.
The discussion was prompted in the following thread:-
https://www.diabetes.co.uk/forum/threads/beta-cell-failure-in-t2.184909/

It was broken out into the following research threads:-
https://www.diabetes.co.uk/forum/threads/interesting-article-on-c-peptide.185124/#post-2461988

https://www.diabetes.co.uk/forum/threads/interesting-report-on-insulin-and-weight-loss.185125/

And resulted in me sending an email to Roy Taylor with some of my concerns.
To avoid lengthy posts I will break these up into discrete steps in following posts.
 
Oldvatr

Oldvatr

Expert
Messages
8,470
Type of diabetes
Type 2
Treatment type
Tablets (oral)
My initial email ran like this
"Dear Sir,
I was in a discussion regarding the DIRECT study results, and a thought came to mind regarding one result in particular.

I was curious in the way the insulin secretion and beta cell function were being presented.
I noted from an interview you gave, which was quoted in the discussion, that the results used Body Surface Area to homogenise the results. (deconvolute?) I.e. to allow cross comparisons between individual results from the cohort.

I can see why this would apply during the insulin clamp trials.

This would explain where the /m^2 term comes from.

But further investigation of BSA shows that it is mathematically derived from the body weight of the participant

I see from another experiment you did that you used a software package (ISEC) to work out the secretion rate during the clamp. The info I have on that software is limited, but it seems to confirm that body weight is indeed used in the calculation, along with height. But it does not identify which of the three formulae it uses for making the calculation.

The error I speak of here is not the results per se that were obtained from the clamp trial. It is the use of the weight term in the equation that concerns me.

While during the clamp trial, neither weight nor height changes significantly, then all the results are valid.

When comparing different results from staged or extended time intervals, such as during the ND trial period and during the Follow On 2-year period, then weight may well change and indeed since the object of the intervention was intended weight loss, then the weight term in the calculation will introduce a confounding influence on the reported results. I think that the weight term ends up as a divider to the c-peptide readings.

So, as the trial progresses, it will introduce a false apparent improvement in beta cell output directly proportional to the weight loss, which I believe is what you reported in the results. It is the other face of the coin. The chicken and the egg.

It is further confirmed by the report of the follow on at 2 years, where you report that those participants who maintained their weight steady did not see any change in insulin secretion. i.e. both c-peptide output AND weight remained steady and unchanged. Some results seem to have weakened, and again this may be the weight increasing post trial rather than beta cell output falling.

I do not have access to any of the raw data or details of the ISEC software, but I propose that the insulin clamp trial does not establish that there was any valid change in beta cell output as a result of the intervention.

I would have expected the units to be pmol/min not pmol/mi/m^2

Yours
"
 
Oldvatr

Oldvatr

Expert
Messages
8,470
Type of diabetes
Type 2
Treatment type
Tablets (oral)
Roy Taylor has replied
"Dear ......,
Thank you for your thoughtful comments.
The use of body surface area is standard for this type of metabolic measurement. Weight loss will indeed decrease surface area although only by a small margin for the weight losses in DiRECT. Using data corrected for weight itself produces a greater effect. However, maximal insulin secretion is only slightly and non-significantly changed at 5 months when the weight loss was maximum (around 15kg). Maximal insulin secretion increased further and became significantly increased only at 12 months at which time weight regain had occurred (to around 12 kg below baseline). Weight increased further to 24 months but the insulin secretion remained steady and sufficient to maintain blood tests in the non-diabetic range.

i hope this is of interest to you and answers your question.
Best,"
 
  • Like
Reactions: ziggy_w
Oldvatr

Oldvatr

Expert
Messages
8,470
Type of diabetes
Type 2
Treatment type
Tablets (oral)
My original email read:-
"Dear Sir,
I was in a discussion regarding the DIRECT study results, and a thought came to mind regarding one result in particular.

I was curious in the way the insulin secretion and beta cell function were being presented.
I noted from an interview you gave, which was quoted in the discussion, that the results used Body Surface Area to homogenise the results. (deconvolute?) I.e. to allow cross comparisons between individual results from the cohort.

I can see why this would apply during the insulin clamp trials.

This would explain where the /m^2 term comes from.


But further investigation of BSA shows that it is mathematically derived from the body weight of the participant

I see from another experiment you did that you used a software package (ISEC) to work out the secretion rate during the clamp. The info I have on that software is limited, but it seems to confirm that body weight is indeed used in the calculation, along with height. But it does not identify which of the three formulae it uses for making the calculation.

The error I speak of here is not the results per se that were obtained from the clamp trial. It is the use of the weight term in the equation that concerns me.

While during the clamp trial, neither weight nor height changes significantly, then all the results are valid.

When comparing different results from staged or extended time intervals, such as during the ND trial period and during the Follow On 2-year period, then weight may well change and indeed since the object of the intervention was intended weight loss, then the weight term in the calculation will introduce a confounding influence on the reported results. I think that the weight term ends up as a divider to the c-peptide readings.

So, as the trial progresses, it will introduce a false apparent improvement in beta cell output directly proportional to the weight loss, which I believe is what you reported in the results. It is the other face of the coin. The chicken and the egg.


It is further confirmed by the report of the follow on at 2 years, where you report that those participants who maintained their weight steady did not see any change in insulin secretion. i.e. both c-peptide output AND weight remained steady and unchanged. Some results seem to have weakened, and again this may be the weight increasing post trial rather than beta cell output falling.

I do not have access to any of the raw data or details of the ISEC software, but I propose that the insulin clamp trial does not establish that there was any valid change in beta cell output as a result of the intervention.

I would have expected the units to be pmol/min not pmol/mi/m^2

Yours Sincerely
"
 
  • Like
Reactions: hankjam
Top Bottom