Shun NagaiCalorie (energy) information shown on pet food packaging is essential for determining appropriate feeding amounts. Feeding neither too much nor too little directly supports the long-term health of dogs and cats, and many countries are moving toward mandatory calorie labeling.
For example, in the United States, revisions to AAFCO (Association of American Feed Control Officials) model regulations have made calorie content (energy value) labeling mandatory for all dog and cat foods. Accurate calorie labeling is indispensable not only for obesity prevention and nutrition management, but also for comparing different foods and for new product development and marketing.
Energy (calorie) calculation formulas for pet food are important tools used to estimate a product’s energy density and to set label values and feeding guidelines. Understanding these formulas is essential for product developers, because the choice and proper use of a formula can significantly affect the accuracy of feeding amounts and the balance of nutrient delivery to dogs and cats.
This article provides a comprehensive explanation of pet food energy calculation formulas for product development professionals. In particular, it focuses on ME (Metabolizable Energy) as the primary indicator—its definition, how it is calculated, comparisons among major calculation models, and key practical points for use in real-world development work.
The big picture of calorie (ME) calculations
First, let’s organize the basic concepts related to energy in pet food. In general, energy values are considered in three stages:
In pet food labeling and feeding design, ME is the primary metric because it best represents the energy dogs and cats can actually utilize. GE and DE may still be referenced in research or ingredient comparisons, but they are generally not shown on retail pet food labels.
AAFCO also standardizes energy declarations as ME, and if a value is estimated by calculation rather than measured by feeding trials, the label must indicate that it is a calculated estimate.
Types of ME (metabolizable energy) calculation formulas
A variety of formulas have been used in the pet food industry to calculate ME. Each differs in method, accuracy, and appropriate applications, with its own strengths and limitations. Below are the major approaches.
General Atwater formula
The Atwater system was originally developed for human foods. It applies physiological fuel values of approximately 4 kcal/g for protein, 9 kcal/g for fat, and 4 kcal/g for carbohydrates (“4-9-4”).
Because these coefficients assume human digestibility, applying them directly to pet food often yields energy values that are too high. In foods with lower digestibility, the resulting value may approach GE and substantially overestimate usable energy. For this reason, the General Atwater formula is rarely used as-is in modern pet food, and the industry generally relies on the Modified Atwater system instead.
Modified Atwater formula
The Modified Atwater method was developed for pet food based on measured data and uses lower coefficients than the General Atwater system. Specifically, it uses 3.5 kcal/g for protein, 8.5 kcal/g for fat, and 3.5 kcal/g for carbohydrates.
An example formula for ME (kcal/kg) is:
ME (kcal/kg) =
[3.5 × (% crude protein) + 8.5 × (% crude fat) + 3.5 × (% carbohydrate as NFE)] × 10
Here, NFE (nitrogen-free extract) is used as a proxy for carbohydrates and is calculated by difference:
NFE (%) = 100 – (crude protein + crude fat + crude fiber + moisture + ash)
For example, if a food contains crude protein 24%, crude fat 15%, and carbohydrate (NFE) 45%, the ME is approximately 3,690 kcal/kg.
The Modified Atwater method is officially recognized by AAFCO, and when feeding trials are not available, it is accepted as an appropriate way to calculate label ME values.
This method is simple and practical, but it can systematically under- or over-estimate ME depending on digestibility. In highly digestible diets, it tends to underestimate ME, while in low-digestibility, high-fiber diets, it can overestimate ME.
Use the Modified Atwater ME Calculator
NRC 2006 formula
In Nutrient Requirements of Dogs and Cats (NRC, 2006), a more accurate ME prediction approach is presented using a multi-step method. It adjusts for digestibility based on crude fiber and uses different urinary energy corrections for dogs and cats. In simplified form, it proceeds as follows:
NRC also provides an alternative version that uses total dietary fiber (TDF) rather than crude fiber (CF) when TDF data are available. Because TDF better captures total fiber (including soluble fiber), it can improve predictive accuracy.
Use the NRC 2006 (CF) ME Calculator
Hall formula
To further increase accuracy, multivariate regression models have been proposed. A well-known example is the equation developed by Hall et al. (2013), which analyzed 558 feeding trial datasets to create new ME prediction formulas.
A key feature is that GE itself (measured) is included as an explanatory variable. An example form for dog foods is:
ME (kcal/kg) =
575 + 0.816 × GE (kcal/kg)
12.08 × crude fat (%)
– 52.76 × crude fiber (%)
– 20.61 × crude protein (%)
– 6.07 × moisture (%)
This is more complex than other methods, but by incorporating measured GE and key compositional factors, it can deliver very high accuracy across diverse formulations.
Hall’s results suggested that the average deviation from measured ME could be reduced to nearly zero. However, due to its complexity and data requirements (including measured GE), it is generally more of an academic model than a practical industry standard at present.
Direct measurement (feeding trials)
All of the above are estimation methods using analytical composition. ME can also be determined directly, following AAFCO feeding trial protocols.
In this method, animals are fed the diet for a defined period and energy intake and energy losses are measured. Fecal and urinary gross energy are measured (e.g., via bomb calorimetry), and ME is calculated as:
GE – fecal GE = DE
DE – urinary GE = ME
A simplified approach may measure only DE and then estimate urinary energy losses (e.g., using a correction per gram of digestible protein).
Direct measurement provides the most accurate ME values but is time-consuming and costly. In practice, most products rely on calculated ME, while feeding trials are reserved for special cases (e.g., therapeutic diets, highly unusual formulas, or novel ingredient evaluations).
How ME calculation works in practice
As described above, pet food energy is ultimately expressed as ME. Here is the basic principle of narrowing energy from “potential” to “usable”:
- Identify energy-providing components
Carbohydrates (NFE) are typically calculated by difference:
NFE (%) = 100 – (crude protein + crude fat + crude fiber + moisture + ash)From there, you determine protein, fat, and carbohydrate levels per 100 g (or on a dry matter basis if needed).
- Calculate GE
NRC approach
Protein = 5.7 kcal/g
Fat = 9.4 kcal/g
Carbohydrates + crude fiber = 4.1 kcal/g - Estimate DE via digestibility
Dogs: Digestibility (%) ≈ 91.2 – 1.43 × crude fiber (% DM)
Cats: Digestibility (%) ≈ 87.9 – 0.88 × crude fiber (% DM) - Calculate ME by subtracting urinary losses
Dogs: ME = DE – (1.04 × % crude protein)
Cats: ME = DE – (0.77 × % crude protein)
In these calculations, crude fiber is a key variable influencing digestibility, and crude protein strongly affects urinary energy loss corrections.
Worked examples and comparisons
Example: Modified Atwater
- Assume a dry food with:
Crude protein: 30%
Crude fat: 15%
Crude fiber: 5%
Moisture: 10%
Ash: 5%
NFE (by difference): 40% - Modified Atwater:
Protein: 30 × 3.5 = 105 kcal/100 g
Fat: 15 × 8.5 = 127.5 kcal/100 g
NFE: 40 × 3.5 = 140 kcal/100 g - Total ME = 372.5 kcal/100 g = 3,725 kcal/kg
This simplicity is the main advantage of Modified Atwater. However, because it does not explicitly adjust for digestibility differences, it can show systematic bias in extreme formulations—underestimating highly digestible diets and overestimating high-fiber, low-digestibility diets.
Use the Modified Atwater ME Calculator
Example: NRC 2006 (dog, CF-based)
Using the same composition:
Step 1: GE
- Protein: 30 × 5.7 = 171 kcal/100 g
- Fat: 15 × 9.4 = 141 kcal/100 g
- NFE + fiber: (40 + 5) × 4.1 = 184.5 kcal/100 g
GE = 496 kcal/100 g
Step 2: Digestibility
Crude fiber on DM basis:
5 / (100 – 10) × 100 ≈ 5.6% (DM)
Digestibility ≈ 91.2 – 1.43 × 5.6
≈ 91.2 – 8.0
≈ 83.2%
Step 3: DE
DE ≈ 496 × 0.832 ≈ 413 kcal/100 g
Step 4: ME
- Urinary loss correction = 1.04 × 30 = 31.2 kcal/100 g
- ME ≈ 413 – 31.2 ≈ 382 kcal/100 g= 3,820 kcal/kg
In this case, the NRC estimate is slightly higher than Modified Atwater.
High-fiber dry dog food comparison
Assume:
NFE: 40%
Crude fiber: 10% (≈11% DM)
Crude protein: 25%
Crude fat: 10%
Moisture: 10%
Modified Atwater
ME ≈ (25×3.5 + 10×8.5 + 40×3.5)×10
= 3,125 kcal/kg
NRC 2006 (dog, CF-based):
GE = (25×5.7) + (10×9.4) + (50×4.1)
= 441.5 kcal/100 g
Digestibility = 91.2 – 1.43×11
= 75.31%
DE = 441.5 × 0.7531
≈ 332.5 kcal/100 g
ME = 332.5 – (1.04×25)
= 332.5 – 26
= 306.5 kcal/100 g
= 3,065 kcal/kg
Here, the difference is about 60 kcal/kg. The key point is that NRC explicitly incorporates fiber-driven digestibility reduction, while Modified Atwater does not.
Use the NRC 2006 (CF) ME Calculator
Fiber and calculation accuracy
Fiber is a major factor affecting ME prediction accuracy. Fiber is less digestible and contributes relatively little energy, so higher fiber generally lowers ME. However, how “fiber” is measured and what type of fiber it is can materially change energy outcomes.
Crude fiber vs total dietary fiber
Crude fiber (CF) typically represents mostly insoluble fiber components captured by traditional chemical analysis, and it tends to miss many soluble fibers (pectins, gums, inulin, etc.). As a result, CF can underestimate total fiber content, especially in diets high in soluble fiber.
Total dietary fiber (TDF) includes both soluble and insoluble fiber and can provide a more realistic measure of total fiber. Because TDF better reflects actual fiber load, prediction formulas using TDF (such as NRC 2006 TDF-based equations) generally perform better.
TDF analysis is more complex and costly, and it is not commonly reported on commercial pet foods. However, when accurate energy prediction is critical—especially for high-fiber formulas—measuring TDF can materially improve estimation quality.
Practical cautions for high-fiber diets
In weight management diets and other high-fiber products, reliance on crude fiber alone can lead to larger estimation errors:
- If a diet contains substantial soluble, fermentable fiber, crude fiber may miss much of it. Some fermentable fibers are partially converted into short-chain fatty acids and can contribute energy. In such cases, calculations based only on crude fiber may underestimate ME.
- If a diet is high in insoluble, poorly fermentable fiber, Modified Atwater can overestimate ME because it does not adjust for digestibility reduction.
FEDIAF has also cautioned that in certain high-fiber dog foods (e.g., where crude fiber on a dry matter basis is high and fermentability is high), prediction formulas may underestimate energy density.
For product developers, the key is to understand the fiber profile of the formulation and choose calculation methods accordingly. If possible, measure TDF and apply NRC TDF-based methods; at minimum, apply conservative assumptions and verify energy values where risk is high.
Practical guide for choosing methods
United States (AAFCO)
In the U.S., calorie content labeling is required. AAFCO recognizes two approaches:
- calculated ME using the Modified Atwater method, and
- measured ME via feeding trials.
Most manufacturers use the calculated approach due to speed and cost. However, because calculated values can deviate in extreme formulations (notably high-fiber diets), developers should consider appropriate safety margins, and potentially conduct measurement trials for special products.
Europe (FEDIAF)
In Europe, FEDIAF recommends methods aligned with NRC 2006, and an EU standard (EN 16967:2017) has established ME prediction equations for dog and cat foods. The standard is generally based on NRC-style approaches and includes notes regarding high-fiber diets.
For products intended for the European market, using the FEDIAF-recommended method as a baseline is typically required for both labeling and feeding guideline design.
Dry vs wet foods
The same principles apply to dry and wet foods, but wet foods have lower energy density due to higher moisture content. As a result, analytical variability can have a larger impact on calculated ME.
In particular, guaranteed analysis values on labels often represent minimums and maximums rather than typical values. Using such label numbers directly for calculation can produce inaccurate results. Developers should use representative “typical” values (average composition data) and, when appropriate, convert to a dry matter basis before calculating.
Labeling and feeding guideline design
Energy formulas are mainly used in two operational areas:
- calorie labeling, and
- feeding guide development.
AAFCO labeling typically requires declaring ME (kcal/kg) and also a common unit such as kcal per cup or per can. If the value is calculated rather than measured, it should be identified as a calculated estimate.
For feeding guidelines, daily energy requirements (DER) are calculated for the target animal and then divided by product ME to determine grams per day. Therefore, if ME is over- or underestimated, feeding recommendations will be inaccurate.
This is especially critical for “light” or low-calorie products, where maximum energy density thresholds may apply depending on claim and category, and where accurate design and labeling are closely tied to regulatory compliance and consumer trust.
Conclusion: calorie calculation is the foundation of product design
This article has explained pet food calorie calculation formulas from fundamentals through major models, calculation steps, and practical points for real-world use. For product developers, the value of mastering these tools is enormous.
Accurate energy calculation is the foundation of product design that supports pet health. By applying the right formula appropriately, developers can tune energy density to the desired target while maintaining balanced nutrition and meeting specific product goals (weight management, growth, senior support, and more).
In manufacturing and quality control, energy calculations also support monitoring of ingredient variability and batch-to-batch consistency, and help guide formula adjustments when needed.
In marketing, energy values grounded in robust calculation and verification provide stronger support for claims such as “lower calorie design” or “high-energy nutrition,” improving credibility with both consumers and professionals.
Finally, it is important to remember that formulas are tools—no single equation is universally perfect. The key is to choose and use methods based on formulation characteristics and the required level of accuracy.
Developers should keep up with scientific and regulatory guidance (from AAFCO, NRC, FEDIAF, and related bodies) and continuously reflect updated knowledge in product design. Deep understanding and practical application of energy calculation formulas are a defining pillar of professionalism in pet food development.