Food energy is measured by burning samples in a device called a bomb calorimeter and then converting the results into the familiar calorie counts on nutrition labels using a set of approximation factors developed over a century ago. The system is more rough-and-ready than most people realize, relying on population-average assumptions about how the human body digests protein, fat, and carbohydrate. Understanding how that number on the package gets there reveals both the ingenuity and the limitations of modern food labeling.
Burning Food to Measure Energy
The most direct way to measure the energy in food is to incinerate it. A bomb calorimeter is a sealed, oxygen-filled steel container submerged in water. A weighed food sample is placed inside, ignited electrically, and burned completely. The heat released raises the temperature of the surrounding water, and that temperature change is used to calculate the total chemical energy in the sample.1Journal of Food and Drug Analysis. Preparation procedures of food and beverage samples for oxygen bomb calorimetry: A scoping review and reporting checklist The result is called “gross energy,” and it represents every last bit of chemical energy the food contains, including portions a human body could never extract.
This is where the gap between lab and life begins. Your body does not combust food in a fireball. It digests food selectively, loses some energy in feces and urine, and spends energy on the digestion process itself. A walnut burned in a calorimeter releases more energy than a walnut passing through your gut, because your body cannot break down every component of the walnut’s cell walls. So the raw bomb-calorimetry number has to be adjusted downward before it can appear on a label.
The Atwater Factors
The bridge between gross energy and the calorie counts you see on food packaging is a set of conversion factors developed in the late 1800s by the American chemist Wilbur Olin Atwater. Rather than burning every individual food product, the system works by measuring the grams of protein, fat, and carbohydrate in a food and multiplying each by a fixed energy value. Those familiar round numbers are 4 kilocalories per gram for protein, 4 for carbohydrate, and 9 for fat.2Food Research International. Review How food structure and processing govern metabolizable energy: a critical review of the food-matrix determinants of energy availability and the limits of the Atwater accounting
These are population averages, not laws of nature. Atwater and his colleagues fed people carefully controlled diets, collected their feces and urine, and burned those as well. By subtracting the energy lost in waste from the gross energy of the food, they arrived at “metabolizable energy,” the portion the body can actually use. The 4-4-9 values are rounded averages of those experiments across mixed Western diets. For any specific food or any specific person, the real number can differ.
A more precise version of the Atwater system uses food-group-specific factors. For example, rice protein is assigned a slightly different value than meat protein, and fruit sugars get a slightly different value than grain starch. Most countries, however, allow or default to the simpler general factors. The calorie count you see on a granola bar is almost certainly 4 times the grams of protein, plus 4 times the grams of carbohydrate, plus 9 times the grams of fat, with adjustments for fiber and alcohol if those are present.
Where the Protein Number Comes From
Measuring protein in food is itself an approximation. Laboratories typically measure nitrogen content, not protein directly. The classic technique involves chemically digesting the food, capturing the nitrogen that is released, and then multiplying by a conversion factor to estimate total protein. The standard conversion factor is 6.25, based on an old assumption that all protein contains 16 percent nitrogen.3Journal of the American Oil Chemists’ Society. Calculation of Nitrogen‐to‐Protein Conversion Factors: A Review with a Focus on Soy Protein
The problem is that different foods have different amino acid profiles, which means their nitrogen content varies. And not all nitrogen in food comes from protein; some comes from other nitrogen-containing compounds. Using 6.25 across the board tends to overestimate protein content for most foods.4PubMed. Converting nitrogen into protein–beyond 6.25 and Jones’ factors There are food-specific correction factors (collectively called Jones’ factors) that try to account for this. Milk proteins, for instance, get a factor of 6.38, while wheat flour gets 5.70. But the default of 6.25 still dominates routine testing, meaning the protein figure on many labels is already slightly inflated before anyone multiplies it by 4 to get calories.
Fiber, Sugar Alcohols, and Alcohol
The simple 4-4-9 framework works reasonably well for a meal of chicken, rice, and olive oil. It gets muddier with ingredients that fall outside those neat categories.
Dietary fiber is technically a carbohydrate, but your small intestine cannot digest most of it. Instead, gut bacteria in your colon ferment some fiber into short-chain fatty acids, which your body can absorb and use for energy.5PubMed Central. Perspective: Physiologic Importance of Short-Chain Fatty Acids from Nondigestible Carbohydrate Fermentation The energy yield from this process is lower than from regular carbohydrate, so many regulatory systems assign fiber a value of about 2 kilocalories per gram rather than 4. Others simply subtract fiber from total carbohydrate before doing the multiplication. The choice of method varies by country, which is one reason the same food can show a different calorie count on labels in different markets.
Sugar alcohols, widely used in sugar-free candy and protein bars, present a similar challenge. They are partially absorbed in the small intestine and partially fermented in the colon, so each one yields a different amount of energy. Sorbitol provides roughly 2.6 kilocalories per gram, maltitol about 2.0, and erythritol essentially zero, because the body absorbs it but cannot metabolize it, excreting it unchanged in urine.6Oxford Academic. Safety of sugar alcohols on human health: a review Labels in the United States often use a blanket estimate or list sugar alcohols separately, leaving you to guess whether the “net carbs” marketing claim on the front actually reflects how much energy your body will extract.
Ethanol, the alcohol in beer and wine, is the other major player outside the standard three macronutrients. It provides about 7 kilocalories per gram. Once consumed, your body oxidizes ethanol first to acetaldehyde and then to acetate, which enters the same energy-producing cycle used by other fuels.7PubMed Central. Ethanol metabolism: The good, the bad, and the ugly The oxidation is irreversible and essentially unregulated, meaning your body keeps processing alcohol regardless of whether it needs the energy right then. Alcoholic beverages are not always required to carry calorie counts on their labels in many countries, so the energy contribution of a few glasses of wine at dinner can be invisible.
How Labels Get Their Numbers in Practice
For most packaged foods, the manufacturer does not run a bomb calorimeter every time a new batch rolls off the production line. Instead, the food is analyzed for its macronutrient composition, and calorie counts are calculated using the Atwater factors. But even measuring those macronutrients is streamlined. Near-infrared spectroscopy has become one of the main routine analytical methods in the food industry, able to measure protein, fat, fiber, and moisture content in seconds by shining infrared light at a sample and analyzing the reflected spectrum.8PubMed Central. The Ability of Near Infrared (NIR) Spectroscopy to Predict Functional Properties in Foods: Challenges and Opportunities
These NIR instruments are calibrated against traditional wet-chemistry methods and bomb calorimetry. Research on cereal products has shown that NIR models can predict gross energy and available energy values within the accuracy required for U.S. nutrition labeling.9PubMed. Rapid prediction of gross energy and utilizable energy in cereal food products using near-infrared reflectance spectroscopy For a factory producing thousands of units a day, this speed matters: a technician can scan samples as they come off the line rather than sending each one to a chemistry lab.
The result is a system where accuracy is layered on approximations. The Atwater factors are themselves averages. The protein figure is estimated from nitrogen. The NIR instrument is calibrated against those older methods. Each layer introduces small errors, but for most commercial foods the total error stays within regulatory limits.
How Accurate Is the Number on the Label?
In the United States, the FDA allows a food product’s actual calorie content to exceed the label value by up to 20 percent. That might sound generous, but there is a logic to it: biological variability between batches of raw ingredients, seasonal variation in crop composition, and measurement uncertainty all create a moving target. A study that measured the actual calorie content of common snack foods found that most products fell within that 20-percent tolerance, though some exceeded it.10PubMed Central. Food Label Accuracy of Common Snack Foods
The Atwater system itself introduces systematic errors in certain food categories. High-fiber diets present a known weakness: the 4-4-9 factors assume a mixed, relatively low-fiber diet, and they tend to overestimate the usable energy in foods with substantial fiber content.11The American Journal of Clinical Nutrition. Accuracy of the Atwater factors and related food energy conversion factors with low-fat, high-fiber diets when energy intake is reduced spontaneously If you eat a lot of vegetables, legumes, and whole grains, the actual metabolizable energy you absorb is probably somewhat less than the label suggests. Conversely, for highly processed foods stripped of fiber, the label may be fairly close to what your body extracts.
Whole almonds are a frequently cited example. Much of their fat is locked inside intact cell walls that resist digestion, so the body absorbs fewer calories from a raw almond than the Atwater calculation would predict. Grinding those same almonds into butter breaks the cells open, making more fat available. The label, however, uses the same calorie figure regardless of form. This discrepancy highlights a fundamental limitation: the system measures what is in the food, not what your body will do with it.
The Cost of Digestion
Even the calories your body does absorb are not all available for storage or activity. Digesting, absorbing, and metabolizing food costs energy, a phenomenon called the thermic effect of food. Evidence indicates this thermic effect is higher for protein and carbohydrate than for dietary fat, and higher for larger meals than for frequent small ones.12PubMed. The Thermic Effect of Food: A Review In rough terms, your body spends about 20 to 30 percent of protein calories just processing the protein, compared to only a few percent for fat. None of this is reflected in the label. Two foods with an identical calorie count can deliver meaningfully different amounts of net energy depending on their macronutrient composition.
This is partly why diet advice about “eating more protein” has a thermodynamic basis beyond satiety. A high-protein meal literally wastes more of its own energy during digestion. The same logic applies to low-fat, plant-based diets, which tend to have a higher thermic effect overall. The calorie label treats all these meals the same, but your body does not.
Food Structure and Processing
The physical structure of food also changes how much energy your body can extract from it. Ultra-processed foods tend to deliver their calories in forms that are already broken down: refined starches, extracted oils, isolated proteins. Research suggests these foods deliver usable calories more rapidly than whole foods, partly because there is less structural work for the gut to do.13PubMed Central. The impact of caloric availability on eating behavior and ultra-processed food reward Your digestive enzymes have immediate access to the nutrients rather than having to break through cell walls, fibrous matrices, and protein networks first.
This speed of calorie delivery has implications beyond simple energy balance. When calories arrive quickly, they may interact differently with gut-brain signaling pathways that regulate appetite. A slice of white bread and a handful of intact wheat berries could have similar label calories per gram, but the bread delivers its glucose faster and with less digestive effort. The Atwater system, designed in an era when most food was minimally processed, was never built to account for these structural differences.
Measuring What Your Body Actually Burns
If measuring the energy in food is one side of the equation, measuring what your body does with that energy is the other. The standard clinical tool for this is indirect calorimetry, which measures oxygen consumption and carbon dioxide production. Because the body’s metabolic reactions consume oxygen and produce carbon dioxide in known ratios depending on whether fat, carbohydrate, or protein is being burned, these gas measurements can be used to calculate both total energy expenditure and which fuel the body is using at any given moment.14International Journal of Obesity. Fasting substrate oxidation at rest assessed by indirect calorimetry: is prior dietary macronutrient level and composition a confounder?
Indirect calorimetry works well in a lab, where you can have someone breathe into a hood or sit in a sealed metabolic chamber. But it cannot follow you around for a week. For measuring energy expenditure in everyday life, researchers use a technique called doubly labeled water. You drink water that contains rare, harmless isotopes of hydrogen and oxygen, then provide urine samples over the following days. The rate at which the isotopes wash out of the body reflects carbon dioxide production, which in turn reflects total energy expenditure.15PubMed Central. Doubly labelled water assessment of energy expenditure: principle, practice, and promise This method has become the gold standard for validating other, cheaper tools. Wearable accelerometers on wrists and thighs, for instance, have been tested against doubly labeled water to see how accurately they estimate daily calorie burn in free-living conditions.16PubMed Central. Estimating energy expenditure from wrist and thigh accelerometry in free-living adults: a doubly labelled water study
The disconnect between the two sides of the equation matters for anyone trying to manage weight. The calorie count on your food is approximate. The calorie count from your fitness tracker is also approximate, calibrated against a method that itself has assumptions. Neither number is wrong, exactly, but treating either as precise is a mistake. The real value of calorie labels is comparative: they are useful for telling you that one food has roughly twice the energy density of another, not for calculating your exact daily intake to the last digit.
Energy Density and Why Volume Matters
Energy density, the number of calories per gram or per bite, turns out to be one of the most practical applications of food-energy measurement. Water contributes zero calories per gram, while fat contributes 9, so the water and fat content of a food largely determine how calorie-dense it is.17PubMed Central. Dietary energy density: Applying behavioural science to weight management A bowl of broth-based soup can fill your stomach while delivering a fraction of the calories in a small handful of nuts.
Research shows that when people are offered foods with lower energy density, they tend to eat fewer total calories without deliberately restricting intake. Adding water-rich ingredients like vegetables and fruits to a meal lowers its overall energy density and can reduce how much energy a person consumes at that sitting. Eating a large-volume, low-energy-dense food like a salad or soup before the main course has a similar effect, enhancing fullness signals and reducing the total calories consumed during the meal.
This is where the abstract science of food-energy measurement becomes genuinely useful for everyday decisions. You do not need to know what a bomb calorimeter looks like or how nitrogen conversion factors work. But understanding that a food’s calorie count reflects its macronutrient composition, that processing increases effective calorie availability, and that volume and water content affect satiety, gives you a more honest picture than the single number on the label ever could.
Why the System Persists Despite Its Flaws
Given all the approximations, it is fair to wonder why regulators have not replaced the Atwater system with something better. The short answer is that no practical alternative exists that works at scale. Measuring the actual metabolizable energy of every food for every person would require feeding studies for each product, accounting for individual differences in gut bacteria, enzyme activity, chewing thoroughness, and cooking method. That is scientifically interesting but commercially impossible.
The Atwater factors persist because they are good enough for their intended purpose: giving consumers a rough tool for comparing foods and managing intake at a population level. They are not meant to tell any individual person exactly how many calories they will absorb from a specific meal. Proposals to update the system generally focus on improving the factors for fiber, sugar alcohols, and whole versus processed foods, rather than abandoning the framework entirely. Some researchers have called for separate factors for different food structures, acknowledging that a calorie from a whole grain and a calorie from refined flour are not metabolically equivalent. Whether regulators will adopt such changes remains an open question, but the core approach of estimating macronutrients and multiplying by fixed values is likely to remain the backbone of nutrition labeling for the foreseeable future.