If you drop fat contents from 60% in high fat to let's say 20% in low fat, obviously the difference is made up by other components (proteins and carbohydrates). Clearly, this will influence various aspects of metabolic activities and this is precisely why one does such studies - to learn direct and indirect effects of nutritional balance.  

If your experimental design includes limited caloric intake, you have to match the calories in two groups and ensure that proteins and other essential components are identical, which leaves you with only carbohydrates for adjustments. On the other hand, if animals are fed ad libitum, caloric intake will most likely be different but that will be part of the design of experiment. If you wish to document the amount of food consumed by animals fed ad libitum, they must be individually housed in metabolic cages. 

You must make them isocaloric unless you are studying effects of differing caloric intake. The metabolic effects of the two diets are vastly different, especially when higher carbohydrate intakes occur--assuming protein intakes remain constant. in humans, the higher fat diets will lower plasma triglycerides. The higher carbohydrate diets will increase triglycerides, lower plasma HDL levels, increase post-prandial glycemic and insulin responses and produce a state resembling many aspects of pre-diabetes, including higher levels of glycated hemoglobin.

Hi, the total amount of calories is the result of the sum of all diet components (carbo, fats, proteins). You may have two different feds with the same amount of calories but heterogenous in terms of composition, which exerts different effects on your results. 

It is highly recommended to adjust proportionally the fed compositions to obtain fair results.

Kind regards

FM

Thanks every one for useful comments.

@ John: Based on your suggestion and since I want to study the effects of a protein hydrolysate on insulin resistance in rats, in order to adjust calories of the 2 diets what is the better approach? only increase carb and keep protein content equal or change them proportionally to adjust the energy? I ask this because my treatment is mainly protein based and I don't know if protein content is also different in 2 diets how that may affect the results and how complicate will be comparison btw groups at the end of the day.

Any suggestions?

Dear Forough: It depends on the study's purpose. First, if seeking evidence for dietary protein being causal in creating insulin resistance is the goal, then you must decide if the quest is for the relative amount (in % of calories) vs the type of protein being fed. Single variable studies are always preferable. Holding then to a single variable, one could, for example, compare casein to other types of protein, with all other dietary components remaining equal in % calories in the two groups. If your goal is to see if a higher fat intake vs a higher carbohydrate intake interacts with protein, with % calories from protein held constant, in causing insulin resistance, then you have a two-variable study, and any result of differences in degree of insulin resistance would be due to an interaction, without knowing a single putative cause.

Forough, I think you could have two experimental groups and two control groups if you have the resources. I would design you experiment based on jWF's response. The percent of protein and the quality of protein should be 'separate' experiments since manipulation of a single variable makes the best experimental design. The problem with studies of insulin resistance and protein is multiple confounders and their potential interactions. I agree with JWF that if your purpose is to assess the interaction of percent of calories from protein with percent of calories from fat and carbohydrate on insulin resistance, then the type and quality of the fat and protein must be held constant between the arms. Other variables to consider are glycemic index, fiber, and genetic variations within and across animal models. 

Thanks again for all valuable comments. I think I should have clearly mentioned the purpose of my study.

Based on the cell culture data we observed potential benefits of a protein hydrolysate and now I'm going to test the hypothesis that this treatment (protein hydrolysate) would ameliorate insulin resistance in the animal model. In literature I see that researchers make rats insulin resistant by feeding high fat diet and that's why I'm going to use this approach. So, I'll feed animals HFD for sometime (until get IR) and then my treatment for several weeks. I'm also going to have a control group (standard diet) to compare to HFD. So I think I should make all groups isocaloric. (the only parameter I would change is carb content to compensate for energy in high fat diet). 

Is this a good practice? 

Peter: Yes, this will be my approach, after induction of IR, HFD will be continuously fed to rats along with my protein hydrolysate (for treatment group) or casein (for control group).

Yes, Peter has a good point. You need a control group of regular chow with protein hydrolysate and regular chow with casein. There would be four groups - HFD with protein hydrolysate; HFD with casein, regular chow with protein hydrolysate; regular chow with casein.

Thanks Joan. Yes, I'll have the 4 groups you mentioned so I can see the effects of my treatment and type of diet on animals.