TLDR: Increasing insulin sensitivity makes you fat.
This study is a bit of a mess because there are no control groups. People either got the safflower intervention or the conjugated linoleic acid intervention, then they were crossed over:
Comparison of dietary conjugated linoleic acid with safflower oil on body composition in obese postmenopausal women with type 2 diabetes mellitus
I'm going to ignore all of the CLA/post CLA data and look at the subjects who got just safflower oil, a total of eight capsules per day, two with each meal plus two at bedtime, eight grams a day of the oil for the first 16 weeks of the study. The safflower oil was 78% linoleic acid, regularly checked by gas chromatography.
Looking at Table 3 there was no change in total fat mass (and subjects didn't gain any weight on the scales) by DEXA scan while there was a loss of 1.2kg of "truncal" adipose tissue. With a PUFA supplement. It appears that DEXA scanning cannot differentiate between visceral and subcutaneous fat in the trunk area. The authors can't quite claim that there was selective loss of visceral fat but I think it is very likely that this did happen.
Throw in a fall in fasting glucose and a downward trend in fasting insulin levels coupled with a rise in adiponectin, some muscle gain and well, that's pretty impressive. You can, absolutely, see why people might have the idea that PUFA could be very positive for metabolic health.
How might one view this from a Protons perspective, other than reaching for a bottle of safflower oil?
I think the first thing to consider is the (probable) loss of visceral fat. Visceral fat, in my opinion, is utterly harmless. It contains the most insulin sensitive adipocytes in the body. If you are chronically hyperinsulinaemic, especially overnight, your insulin may never drop low enough to release any significant lipid from your visceral fat. So visceral fat is a surrogate for nocturnal hyperinsulinaemia, which is what is actually bad for you.
We have values for 10h fasting insulin; at enrolment it was 19.9microU/ml and this dropped to 18.2microU/ml over the first 16 weeks of the study. I would not expect 19.9microU/ml to maintain visceral fat and 18.2microU/ml to melt it away. I think it is much more likely that the gross hyperinsulinaemia induced by the sort of evening meal recommended by the ADA for diabetic people might well have resolved faster with safflower oil supplementation than it did without safflower oil, ie the duration of the period of gross hyperinsulinaemia through the night was reduced. Fasting levels were unchanged but the time spent above this ought to have been reduced.
We just have to revisit the Spanish study to see why:
This graph is over eight hours, 10 hours would be similar. These are healthy volunteers, the hyperinsulinaemia would be worse in DMT2 patients eating a high carbohydrate meal. Black squares are butter, white triangles are a high PUFA seed oil. The higher the PUFA content of the meal, the faster insulin level drops. Adding PUFA a mixed meal should allow insulin to drop faster and sooner than saturated fats. This happens because PUFA fail to generate the ROS needed to maintain the physiological insulin resistance which ought to occur post prandially to limit calorie ingress in to cells, adipocytes included. This leaves glucose and fatty acids available to signal satiety to the brain. Also noted in the Spanish study was that PUFA induced more rapid clearance of chylomicrons and more rapid drop in FFAs compared to saturated fats. As I wondered at the time, where do the FFAs and chylomicrons go to?
They go in to adipocytes, because the adipocytes cannot say "no" if PUFA generate too little ROS.
So this drug (safflower oil) allows increased insulin sensitivity (reminiscent of the "glitazones") or, rather, it fails to generate the ROS needed to limit the over expansion of adipocytes, which shows as increased insulin sensitivity during peak insulin exposure. This increased insulin sensitivity puts calories in to adipocytes rapidly so reduces the need for sustained hyperinsulinaemia. All adipocytes gain fat, but the faster fall in insulin allows an increase in the time window where visceral fat can actually release at least some FFAs to the systemic circulation via the portal vein and liver. Visceral fat shrinks, non-visceral fat expands.
The "benefit" of reducing visceral fat in this way during fasting is paid for by increasing the non-visceral fat depots in the immediate post prandial period. The extra fat in non-visceral adipose tissue will come primarily from the diet and the lost fat from visceral adipocytes will be used to provide fasting calories. In this particular study, the amount gained by non-visceral adipocytes was roughly equal to that lost by visceral adipocytes, it's probably random chance that the numbers balanced. And DEXA seems a pretty crude technology to use to work in small numbers of grams of adipose tissue, just looking at the non-balancing cited changes in fat and lean tissue mass in the results.
These processes can continue until non-visceral fat mass eventually become high enough that the loss of FFAs due to adipocyte distension over rides the insulin sensitising effect of the safflower oil. At this point overall insulin exposure will increase and visceral fat will return, on top of a higher mass of non-visceral adipose tissue. It will take longer than 16 weeks.
If you are an obese diabetic taking part in a study like this you should see a prompt but transient improvement in insulin sensitivity. This enhanced sensitivity should allow more non-visceral fat gain until you convert to being a somewhat more obese diabetic. It nicely illustrates that extra PUFA convert you from being established "obese" to being "pre-more-obese". Time is all that is needed to convert you from being "pre-more-obese" to simply"more-obese".
But your lab numbers will improve transiently in the first part of the intervention.