Simplified from CALCULATION OF THE ENERGY CONTENT OF FOODS - ENERGY CONVERSION FACTORS (FAO.org)
Food energy can be:
- Combustible or ingested energy = theoretical maximum energy content of a food measured using bomb calorimetry (the energy you ingest)
- Metabolizable energy (ME) = ingested energy minus energy lost in feces by indigestible nutrients (the energy listed on nutrition labels)
- Net metabolizable energy (NME) = metabolizable energy minus energy converted into heat due to dietary-induced thermogenesis
Comparison of ME (from food labels) and NME (potentially fattening energy) in different macronutrients (FAO.org, table 3.3.):
- Protein (Calories/gram): ME = 4, NME = 3.2
- Fat: ME = 9, NME = 9
- Carbohydrates: ME = 4, NME = 4
- Dietary fiber: ME = 2 (but wrongly counted as 4 on food labels), NME = 1.4
- Alcohol: ME = 7, NME = 6.3
On food labels, dietary fiber is listed under carbohydrates as having 4 Cal/g, but its ME is only 2 Cal/g and NME only 1.4 Cal/g. So, a certain carbohydrate food that has a lot of fiber can have significantly less energy than stated on the food label.
The metabolic energy (ME) of individual amino acids in proteins can vary from 2 to 6 Cal/g (ResearchGate). Most proteins contain most amino acids, but in different proportions. I haven't found, so far, if this results in significantly different caloric value of various proteins, such as in beef and egg white protein, for example.
In conclusion, net metabolic energy (NME), which is potentially fattening energy, tends to be lower than metabolic energy (ME) stated on the food labels, at least for proteins, dietary fiber and alcohol.
The NME of certain nutrients can further differ due to personal factors, such as age, state of health, etc. and nutrient combinations (e.g., fiber can slightly inhibit the absorption of fat) (European Journal of Clinical Nutrition). So, Calories stated on the food labels are not fixed values, but, so far, I haven't found any evidence how could be this practically important.