Dietary Reference Intakes (DRIs) are a set of scientifically derived reference values used to assess and plan nutrient and energy intakes for generally healthy people. In their specific institutional sense, they are the framework developed jointly for the United States and Canada through the National Academies of Sciences, Engineering, and Medicine and its predecessor organizations. DRIs distinguish average requirements, intakes sufficient for nearly all individuals, upper intake limits, and other measures. They are foundational tools in nutrition research, food policy, and public health, rather than a single recommended intake or a diagnostic standard.
Scope and development
DRIs describe usual intake over time, not the amount that must be consumed on every individual day. Values are organized by age and sex, with separate categories where appropriate for pregnancy and lactation. They cover vitamins, mineral nutrients, protein, other macronutrients, dietary fiber, water, and dietary energy.
The framework expanded the earlier Recommended Dietary Allowances used in the United States and Recommended Nutrient Intakes used in Canada. The first United States Recommended Dietary Allowances were issued in 1941. The joint DRI series began publishing nutrient reports in 1997, extending the emphasis beyond preventing deficiency to include excessive intake and, where evidence permits, chronic disease risk.
DRIs are revised through separate reports rather than a single regularly replaced master document. Consequently, reference values for different nutrients may reflect assessments completed in different years.
Outside the United States and Canada, other authorities use related frameworks, including dietary reference values and nutrient reference values. Their terminology, age groupings, assumptions, and numerical values are not necessarily identical.
Main reference values
The DRI framework contains several complementary measures. Their names describe different statistical or biological purposes.
| Reference value | Meaning | Principal interpretation |
|---|---|---|
| **[[estimated-average-requirement | Estimated Average Requirement (EAR)]]** | Intake estimated to meet the requirement of half the healthy individuals in a defined life-stage and sex group |
| **[[recommended-dietary-allowance | Recommended Dietary Allowance (RDA)]]** | Average daily intake sufficient to meet the requirements of nearly all—about 97–98%—healthy individuals in a group |
| **[[adequate-intake | Adequate Intake (AI)]]** | Intake assumed to be adequate, based on observed or experimentally determined estimates in apparently healthy groups |
| **[[tolerable-upper-intake-level | Tolerable Upper Intake Level (UL)]]** | Highest average daily intake likely to pose no risk of adverse health effects to almost all people in the general population |
| **[[estimated-energy-requirement | Estimated Energy Requirement (EER)]]** | Average dietary energy intake predicted to maintain energy balance, with additional requirements for growth or relevant physiological states |
| **[[acceptable-macronutrient-distribution-range | Acceptable Macronutrient Distribution Range (AMDR)]]** | Range of intake for an energy source, expressed as a percentage of total energy |
| **[[chronic-disease-risk-reduction-intake | Chronic Disease Risk Reduction Intake (CDRR)]]** | Intake above which reduction is expected to reduce chronic disease risk in an apparently healthy population |
Not every nutrient has every type of value. An AI ordinarily substitutes for an EAR and RDA rather than supplementing them. A UL or CDRR may be absent because the evidence does not support establishing one.
EAR and RDA
The EAR concerns the middle of the requirement distribution; the RDA concerns its upper end. Where requirements are approximately normally distributed and their variability can be estimated, the RDA is commonly calculated as:
Here, represents variation in requirements between individuals, not variation in one person’s daily food intake. Other methods are used when the requirement distribution or available evidence does not support this calculation.
An intake below the RDA does not by itself establish inadequacy. The RDA deliberately exceeds the requirements of most people in its reference group.
Adequate Intake
An AI is established when evidence cannot support a sufficiently reliable requirement distribution. It may draw on observed intakes in populations judged to have adequate nutritional status, experimental evidence, or extrapolation.
An AI is therefore not an EAR with a different name. Usual intake at or above an AI generally supports a judgment of adequacy, but intake below it does not provide a quantitative estimate of the probability of inadequacy.
Upper limits and chronic disease reduction
A UL is not a recommended intake, and the absence of a UL does not demonstrate that unrestricted intake is safe. Intake above a UL indicates increasing potential for adverse effects, not certainty that harm will occur.
The intake sources covered by a UL depend on the nutrient and the adverse effect used to derive it. Some ULs apply to total intake from food, beverages, and supplements; others apply only to supplemental or added forms.
The CDRR separates evidence about chronic disease reduction from the assessment of toxicity used for a UL. The 2019 sodium and potassium report introduced a sodium CDRR. For adults, it identified 2,300 milligrams per day as the intake above which reduction is expected to reduce chronic disease risk. This is not equivalent to defining all intakes above that amount as toxic.
Scientific basis
DRI committees evaluate evidence on nutrient intake, physiological requirements, health outcomes, and adverse effects. Relevant evidence can include controlled feeding studies, randomized controlled trials, observational studies, balance studies, and measurements of nutritional status.
The selected adequacy criterion is important. A requirement might be defined by maintaining body stores, supporting normal function, or achieving an appropriate biomarker value. Different criteria can yield different estimates.
Other considerations include:
- Absorption and utilization: Bioavailability can differ by food source, chemical form, and the composition of the diet.
- Life stage: Growth, pregnancy, lactation, and aging can change requirements or the interpretation of available evidence.
- Variation between individuals: Requirements differ even among healthy people of the same age and sex.
- Evidence gaps: Values for some age groups depend partly on extrapolation rather than direct experimental evidence.
- Uncertainty about harm: Upper limits incorporate judgments about the strength of evidence and uncertainty in translating study findings to the wider population.
Energy differs from many nutrients because a sustained intake above an individual’s expenditure generally produces energy storage rather than an additional margin of adequacy. EER equations therefore estimate energy requirements rather than adding a population-wide allowance analogous to an RDA.
Assessment and planning
Individual assessment
DRIs provide a framework for interpreting usual intake, but a dietary record alone does not establish nutritional status. Reported intake is affected by day-to-day variation, reporting error, food-composition estimates, and uncertainty about an individual’s actual requirement.
Nutritional diagnosis may also depend on clinical findings, biochemical measurements, absorption, and disease-related changes in metabolism. DRIs primarily describe generally healthy populations and do not define therapeutic requirements for people with conditions such as chronic kidney disease.
Population assessment
For many nutrients, the proportion of a group whose usual intake falls below the EAR can estimate the prevalence of inadequate intake. This EAR cut-point method depends on assumptions about the requirement distribution and its relationship to intake.
It is not appropriate in every case. Iron requirements in menstruating populations are markedly skewed, so a probability-based approach is generally used instead. Energy inadequacy also cannot be estimated simply by counting reported intakes below the EER, because energy intake and energy requirements are related.
The proportion below the RDA is not an estimate of deficiency prevalence. Similarly, when only an AI exists, the proportion below it cannot be interpreted as the prevalence of inadequate intake.
Food policy and institutional use
DRIs inform food and nutrition surveys, institutional menu planning, research, and evaluations of food fortification and dietary supplement exposure. They also provide part of the scientific basis for dietary guidance.
For group planning, the objective is to shape an intake distribution with a low prevalence of inadequacy and a low prevalence of excessive intake. Merely setting the group’s average intake equal to an RDA does not establish that this objective has been achieved.
Distinction from labels and food-based guidelines
DRIs are not interchangeable with the Daily Values used on United States food labels. Labeling values provide standardized comparison points, often across broad population categories, whereas DRIs distinguish multiple life-stage and sex groups and serve several different assessment purposes. A nutrient’s percentage Daily Value therefore does not directly express the percentage of a particular person’s requirement supplied by a food.
DRIs also differ from food-based dietary guidelines. They describe nutrient and energy reference levels; food-based guidelines translate broader evidence into dietary patterns and food-group guidance. Meeting numerical nutrient reference values alone does not fully characterize dietary quality.
Limitations
DRIs are population-based estimates, not precise measurements of every individual’s needs. Their interpretation is constrained by uncertainty in intake measurement, limited long-term experiments, incomplete evidence for some life stages, and interactions among nutrients.
Evidence for chronic disease outcomes can be especially difficult to interpret because long-term dietary exposures are hard to measure and often correlate with other behaviors. Reference values may consequently change when stronger evidence becomes available or when the endpoint used to define adequacy changes.