Trends in Overall and Micronutrient-Containing Dietary Supplement Use in US Adults and Children, NHANES 2007–2018

ABSTRACT

Background

Dietary supplement (DS) use is widespread in the United States and contributes large amounts of micronutrients to users. Most studies have relied on data from 1 assessment method to characterize the prevalence of DS use. Combining multiple methods enhances the ability to capture nutrient exposures from DSs and examine trends over time.

Objectives

The objective of this study was to characterize DS use and examine trends in any DS as well as micronutrient-containing (MN) DS use in a nationally representative sample of the US population (≥1 y) from the 2007–2018 NHANES using a combined approach.

Methods

NHANES obtains an in-home inventory with a frequency-based dietary supplement and prescription medicine questionnaire (DSMQ), and two 24-h dietary recalls (24HRs). Trends in the prevalence of use and selected types of products used were estimated for the population and by sex, age, race/Hispanic origin, family income [poverty-to-income ratio (PIR)], and household food security (food-secure vs. food-insecure) using the DSMQ or ≥ 1 24HR. Linear trends were tested using orthogonal polynomials (significance set at P < 0.05).

Results

DS use increased from 50% in 2007 to 56% in 2018 (P = 0.001); use of MN products increased from 46% to 49% (P = 0.03), and single-nutrient DS (e.g., magnesium, vitamins B-12 and D) use also increased (all P < 0.001). In contrast, multivitamin-mineral use decreased (70% to 56%; P < 0.001). In adults (≥19 y), any (54% to 61%) and MN (49% to 54%) DS use increased, especially in men, non-Hispanic blacks and Hispanics, and low-income adults (PIR ≤130%). In children (1–18 y), any DS use remained stable (∼38%), as did MN use, except for food-insecure children, whose use increased from 24% to 31% over the decade (P = 0.03).

Conclusions

The prevalence of any and MN DS use increased over time in the United States. This may be partially attributed to increased use of single-nutrient products. Population subgroups differed in their DS use.

Introduction

The use of dietary supplements (DSs) is widespread in the United States and has steadily increased over the last several decades (1–4). Over one-half of US adults and one-third of US children use ≥1 products (1, 5–7), with use especially high in certain population subgroups, such as older adults (8, 9). According to the Nutrition Business Journal's recent Supplement Business Report, supplement sales increased by 14.5% in 2020 compared with 2019, and are projected to grow by US1.8 billion each year (10). As of 2017–2018, multivitamin-mineral (MVM) products, which provide >100% of the recommended daily value for many nutrients, remain the most commonly used DSs in the United States (11, 12). The effect of supplementation on chronic disease risk continues to be debated (13, 14). However, sufficient nutrient intakes are important for health (15, 16), and DSs contribute substantial amounts of essential micronutrients (17–19). Monitoring trends in DS use in the US population over time is important for understanding temporally the role, status, and contribution of DSs (e.g., in terms of total nutrient intake) to the nutrition of the US population (20).

Accurate and reliable dietary assessment methods for assessing nutrient intakes from foods, beverages, and DSs are critical for population-level research and public health monitoring. The strengths and limitations of frequency-based methods and 24-h dietary recalls (24HRs) for foods and beverages are known (21–23); however, much less is known regarding these methods in terms of measuring the prevalence of use of and estimating nutrient intakes from DSs (17). The reference periods of these instruments also differ (i.e., short-term compared with long-term), impacting how the prevalence of DS use is defined, and making comparisons of estimates across studies difficult, if not impossible (17). A small but growing literature on assessment methods for DSs suggests that combining multiple methods of dietary supplement assessment enhances the ability to capture nutrient exposures from DSs (24–26). Extensive research on trends in the prevalence of DS use in US children and adults has been conducted previously (1, 7, 27, 28); however, most if not all of these studies relied on DS data collection via a single dietary assessment method alone, and these methods are often prone to measurement error. Therefore, additional studies updating prevalence estimates of DS use using multiple methods of dietary assessment are needed. Also needed is a well-described classification system for DS products.

The objective of this analysis was to characterize DS use and examine trends in overall and micronutrient-containing (MN) DS use in the US population (≥1 y) and in population subgroups from 2007 to 2018, estimated from national surveys using multiple modes of DS assessment. Population subgroups included those defined by sex, age, race and Hispanic origin, family income, and household food security status.

Methods

NHANES study design

The NHANES is a nationally representative, continuous cross-sectional survey of the noninstitutionalized, civilian US population, conducted by the CDC National Center for Health Statistics (NCHS) (29). NHANES employs a complex, stratified, multistage probability cluster sampling design, and releases data as publicly available 2-y datasets. Details regarding the survey design are described elsewhere (30–32). Briefly, the NHANES protocol includes data collection at 3 time points: during an in-home interview, during a standardized health examination in the mobile examination center (MEC), and during a follow-up dietary interview conducted via telephone. Written informed consent was obtained for all participants or proxies; the NHANES protocol (and publicly released deidentified data) was approved by the Research Ethics Review Board at NCHS.

For the purposes of this analysis, demographic and dietary data from participants in 6 NHANES cycles (2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, and 2017–2018) with complete DS information reported on two 24HRs and the in-home interview were combined to form a sample of 59,842 participants to evaluate trends in DS use over a 10-y span. Those aged <1 y, had incomplete or unreliable 24HR or Dietary Supplement and Prescription Medicine Questionnaire (DSMQ) data, or who were pregnant and/or lactating were excluded, resulting in a final analytic sample of 49,364 US adults and children (≥1 y).

DS assessment

A DSMQ inclusive of an in-home product inventory was administered during the NHANES household interview to assess a participant's DS use in the previous 30 d. Participants were queried on their use of vitamins, minerals, and non–micronutrient-containing (non-MN) DSs, both over the counter and prescription, and were asked to show trained interviewers the containers for all products taken over the past 30 d. Respondents included: a proxy for participants <16 y of age (except emancipated minors) and those who were unable to answer the questions themselves; and participants ≥16 y of age (and emancipated minors) who were able to answer the questions themselves. For each DS reported, the interviewer recorded the name, manufacturer, product form (e.g., tablet), and dose per serving from the label (for single-nutrient products). Additional details on the consumption frequency in the past 30 d, amount usually taken, and duration of DS use were also collected. If product containers were not available, participants were probed to recall in detail the product they had taken. Trained nutritionists at NCHS then matched products reported by participants to product labels according to the information provided in order to gain knowledge of the ingredients and amount per serving. Further information regarding the NHANES DS component protocol is available elsewhere (33–38).

Following the completion of the in-home interview, two 24HRs were collected by trained interviewers using the USDA's automated multiple-pass method (39, 40). The first 24HR was administered in-person in the MEC, followed by a second 24HR completed via telephone ∼3–10 d later. During both 24HRs, participants reported DSs consumed from midnight to midnight the previous day. Respondents for the 24HRs were as follows: a proxy for participants <6 y of age; a proxy with the assistance of the child aged 6 to 8 y; participants aged 9–11 y with the assistance of a proxy; and participants aged ≥12 y who answered by themselves.

DS nutrient values were computed using the NHANES Dietary Supplement Database (41, 42). DS users were identified based on whether participants responded “yes” to taking any products on the DSMQ or at least one 24HR. Specifically, if the participant responded to the question, “Have you used or taken any vitamins, minerals or other dietary supplements in the past 30 days?” on the DSMQ, only then were the DSMQ intakes used. If DS intakes were not reported on the DSMQ in-home inventory, but the participant responded to the question “Any dietary supplements taken in the past 24 hours?” on at least one 24HR, then the intakes reported on at least one 24HR were used. Mean (i.e., average) daily DS intakes were calculated using the proportion of reported days of DS use over the past 30 d, multiplied by the amount the participant reported taking per day, if DS intakes were reported on the DSMQ. If DS intakes were not reported on the DSMQ or at least one 24HR, then the mean nutrient intake from supplemental sources reported on the 24HRs from day 1 and day 2 was used. Nutrient intakes from supplemental sources for vitamins A and E were not summarized in the publicly available NHANES 2007–2018 DS data files; therefore, mean daily intakes from supplements were not estimated for these nutrients.

DS product type classification

Exclusive product type categories of DSs were constructed to examine DS use among users, including MN and non-MN product types. At the product level, the focus of our analysis was predominantly tailored toward MN DSs because these products contain nutrients that can aid in mitigating nutrient insufficiencies, and they are the most commonly used DSs. However, the major non-MN DS product categories were also included for selected analyses in the present study because their use has recently increased in certain population subgroups (27, 43). MN product types included MVM, multivitamin (MV), multimineral (MM), calcium and vitamin D, single vitamin (i.e., thiamin, riboflavin, niacin, folic acid, and vitamins A, B-6, B-12, C, D, and E), and single mineral (i.e., calcium, iron, magnesium, potassium, selenium, and zinc) products. These products may or may not contain non-MN ingredients, such as herbals or botanicals, but are vitamin and/or mineral products that are predominantly MN. Non-MN product types included omega-3 (ω-3; n–3) fatty acids/fish oil, herbal/botanical, probiotic, fiber, coenzyme Q-10, joint, protein/sport, Echinacea, and melatonin supplements. Prevalence estimates of these broad, mutually exclusive supplement categories were adapted from Bailey et al. (44) and Gahche et al. (8), and were constructed based on their nutrient content and/or descriptive characteristics commonly used in marketing them. Types of products were categorized according to the product classification system described in Table 1. Subcategories were constructed to allow for more specific analyses of single-vitamin and single-mineral products. The specific MN and non-MN product types included in this analysis were chosen according to their high prevalence of use and/or importance to nutritional status in the US population. Prevalence estimates for specific single-vitamin and single-mineral product subcategories in DS users were reported; however, estimates with an overall prevalence of <1% in DS users are reported separately in Supplemental Table 1 and were not evaluated for mean daily nutrient intake from supplements. MVMs were defined as products containing ≥3 vitamins and ≥1 mineral. MVs were defined as vitamin combinations containing ≥2 vitamins without minerals, and MMs were defined as mineral combinations containing ≥2 minerals without vitamins. Calcium and vitamin D DSs were defined as any product containing both calcium and vitamin D as the primary ingredients. Single vitamin or mineral DSs were defined based on whether the DS contained a single vitamin or mineral, respectively, without the inclusion of any other nutrients. Certain product categories included in these definitions are not based on the NHANES ingredient count, but rather according to the supplement name in NHANES (i.e., NHANES variable “DSDSUPP”), nutrient content, and/or descriptive characteristics. Non-MN product categories were constructed according to primary ingredient information, the supplement name in NHANES, and/or descriptive characteristics. Antacid products are legally classified as over-the-counter medications rather than DSs, and they do not commonly follow the same usage patterns as DSs; therefore, these products were excluded from all analyses.

TABLE 1.

Classification system used to create dietary supplement product categories, NHANES 2007–2018¹

Mutually exclusive product category	Subcategory	Definition	Examples
Multivitamin-mineral (MVM)		Any product containing ≥3 vitamins and ≥1 minerals; may or may not contain herbals or botanicals	Centrum Silver
Multivitamin (MV)		Any product containing ≥2 vitamins without minerals; may or may not contain herbals or botanicals	B complex
Multimineral (MM)		Any product containing ≥2 minerals without vitamins; may or may not contain herbals or botanicals	Magnesium and zinc
Calcium and vitamin D		Any product containing both calcium and vitamin D as the primary ingredients, may or may not contain other vitamins or minerals; not part of an MVM	Caltrate Bone Health
Single vitamin2	Vitamin A Thiamin Riboflavin Niacin Vitamin B-6 Folate, DFE Folic acid Vitamin B-12 Vitamin C Vitamin D Vitamin E	Any product that contains a single vitamin count of the vitamin of interest (e.g., vitamin A) without other nutrients	Vitamin A single-nutrient DS
Single mineral2	Calcium Iron Magnesium Potassium Zinc	Any product that contains a single mineral count of the mineral of interest (e.g., calcium) without other nutrients	Calcium single-nutrient DS
Botanicals/herbals3		Any product with a botanical ingredient count >1 without vitamins or minerals	Gingko biloba, St John's wort
Omega-3/fish oil3		Any product with omega-3 fatty acids (i.e., EPA, fish oil, flaxseed) as the primary ingredient	Fish oils, DHA, flax seed
Probiotics3		Any product with prebiotics or probiotics as the primary ingredient	Align Probiotic
Fiber3		Any product with fiber as the primary ingredient or laxative in the name	Metamucil reported as a supplement
Coenzyme-Q103		Any product with coenzyme-Q10 as the primary ingredient	Nature's Bounty Co Q-10
Glucosamine, chondroitin (joint supplements)3		Any product containing glucosamine, chondroitin, a combination, or MSM	Glucosamine and chondroitin, Joint Juice, MSM
Protein and amino acids (protein/sports)3		Any product that contains protein, amino acids, or is intended to enhance athletic performance	Creatine, Hydroxycut, lysine, arginine
Echinacea3		Any product that contains Echinacea as the primary ingredient	Puritan's Pride Echinacea
Melatonin3		Any product that contains melatonin as the primary ingredient	Nature's Bounty Melatonin

¹Vitamin counts are identified by the NHANES variable “DSDCNTV”; mineral counts are identified by the NHANES variable “DSDCNTM.” DFE, dietary folate equivalent; DS, dietary supplement; MSM, Methylsulfonylmethane.

²Certain product categories included in this definition are not based on the NHANES ingredient count, but rather according to the supplement name in NHANES (i.e., NHANES variable “DSDSUPP”), nutrient content, and/or descriptive characteristics commonly used in marketing.

³Certain products included in this category may be classified according to the supplement name in NHANES (i.e., NHANES variable “DSDSUPP”) and/or descriptive characteristics commonly used in marketing, in addition to information regarding the primary ingredient.

Demographic and lifestyle characteristics

Across all NHANES cycles, demographic and lifestyle data on age, sex, race and Hispanic origin, educational attainment, health insurance coverage, family income, household food security, smoking status, alcohol use, self-reported health status, and screen time were collected by interviewers using the Computer-Assisted Personal Interview system during the in-home interview. Age was categorized according to the DRI age groupings: 1–3 y, 4–8 y, 9–13 y, 14–18 y, 19–30 y, 31–50 y, 51–70 y, and >70 y. Race and Hispanic origin was categorized by NCHS as non-Hispanic white, non-Hispanic black, and Hispanic or Mexican American. Beginning in 2011, NHANES began oversampling non-Hispanic Asian Americans; however, due to lack of data availability in NHANES 2007–2010, this population subgroup was not included in analyses stratified by race and Hispanic origin for the present study. Similarly, the “other race” category included in NHANES is not presented separately herein but is included in the overall prevalence estimates. Health insurance coverage of the participant at the time of the survey was categorized as either public, private (including those covered under both private and public plans), or uninsured (8).

Income was classified according to the family poverty-to-income ratio (PIR), which relates family income to family size (45). PIR has been used as an indicator of family income level and as an eligibility criterion for federal nutrition assistance programs. Specifically, a PIR of 130% indicates potential eligibility for the Supplemental Nutrition Assistance Program that provides cash benefits for foods to low-income households to reduce food insecurity (46). A PIR of 350% has also been used in previous studies to differentiate middle-income and high-income families (19, 47). Thus, 3 PIR categories were used in this analysis: ≤130%, 131% to 350%, and >350%. Household food security over the previous 12 mo was assessed during the in-home interview via the USDA's Household Food Security Survey Module (48, 49); households responded to either 18 (households with children) or 10 total questions (households without children) and were categorized into 1 of 4 food security classifications accordingly (full, marginal, low, very low). Households considered to have full or marginal food security were combined and considered to be food secure (<3 affirmative responses); those with low or very low food security were combined and considered to be food insecure (≥3 affirmative responses) (48, 49).

Current smoking status, alcohol use, educational attainment, and self-reported health status were evaluated in US adults (≥19 y) only. Current smoking status was determined based on whether participants were never smokers (smoked <100 cigarettes per lifetime), former smokers (>100 cigarettes per lifetime but do not currently smoke), or current smokers. Current smokers were then further classified based on whether they smoked cigarettes daily (current, daily) or whether they classified themselves as a smoker, but did not smoke cigarettes daily (current, occasional) (50). Alcohol consumption was assessed using the NHANES Alcohol Use Questionnaire, which measured use in the last 12 mo, frequency, and number of drinks consumed, and was categorized as 0, 1, 2, or ≥3 drinks/d, consistent with previous literature (8, 51). Educational attainment was categorized as less than high school, high school diploma or general equivalency diploma, some college (or associates degree), or more than college. Current self-reported health status was classified as excellent/very good, good, or fair/poor.

Screen time was calculated as the sum of the total time spent looking at a television and/or computer screen per day for those aged 2 to 18 y (n = 14,454 with complete data) using the NHANES Physical Activity Questionnaire. However, it is important to note that screen time data were only available for children aged ≥2 to 11 y in NHANES 2007–2010, therefore restricting our ability to evaluate screen time in those aged 12–18 y in NHANES 2007–2010. The response of “<1 h/d” was assigned 0.5 h/d as recommended (52); and screen time was categorized as follows: ≤1 h/d, >1–2 h/d,  >2–4 h/d, and >4 h/d. The American Academy of Pediatrics recommends limiting leisure screen time to ≤2 h/d (53).

Weight status, defined by BMI (kg/m²), was obtained from height and weight measured during the health examination in the MEC for adults (≥19 y). BMI classifications were as follows: underweight (<18.5), normal weight (18.5–24.9), overweight (25.0–29.9), and obese (≥30) (54).

Statistical analysis

All statistical analyses were performed using SAS (version 9.4; SAS Institute Inc) and SAS-callable SUDAAN (version 11; RTI International) software programs to account for the complex survey design. NHANES interview sampling weights were used for all analyses to adjust for differential nonresponse and noncoverage, and planned oversampling of some population subgroups and poststratification. Descriptive statistics, such as prevalence estimates (overall and across NHANES cycles) and estimated mean daily nutrient intake from DSs, were estimated using proc descript in SUDAAN, and SEs for all statistics of interest were approximated using Taylor series linearization. The statistical reliability of estimates was determined based on the relative SE as recommended by NCHS (55); values presented have a relative SE ≤30% unless otherwise stated. For demographic and lifestyle characteristics, differences between groups were determined via Satterthwaite-adjusted Wald χ² tests for categorical variables. Linear trends in DS use across NHANES cycles were tested using orthogonal polynomials contrasts. A P value <0.05 was considered statistically significant, but because the potential for type I error exists, exact P values were presented to enhance interpretation.

Results

Characteristics associated with DS use

In 2007–2018, over half (53%) of the US population (≥1 y) reported taking ≥1 DS/d, and these were predominantly MN DSs (Table 2). The prevalence of any and MN DS use was significantly higher in women and girls (58%), older adults (81%), and non-Hispanic whites (59%), as well as those with a higher family income (63%), greater food security (56%), and private health insurance (59%). Among adults (≥19 y), those who were at a normal weight (59%), former smokers (67%), consumed 1 alcoholic drink per day (67%), reported having excellent or very good health status (62%), and those who had a college-degree level educational attainment (69%) were significantly more likely to take a DS. In children aged 2 to 11 y, ∼50% of those who reported participating in fewer hours of screen time (<2 h) per day took a DS, whereas only 39% of those participating in >4h/d did so (Table 2).

TABLE 2.

Prevalence of overall and micronutrient-containing supplement use by demographic and lifestyle characteristics in the US population (≥1 y), NHANES 2007–2018¹

	Total sample		Reported on DSMQ or at least one 24HR
		Any DS			MN DS
Characteristics	n	n	% (SE)	P value2	n	% (SE)	P value2
% of total population	49,364	23,050	52.8 (0.6)		20,950	48.0 (0.5)
Sex of children (1–18 y)				0.2302			0.1621
Boys	9315	3092	37.1 (1.0)		2843	34.0 (0.9)
Girls	8996	3018	38.3 (0.9)		2805	35.4 (0.9)
Sex of adults (≥19 y)				<0.0001			<0.0001
Men	15,403	7052	51.2 (0.7)		6564	44.9 (0.7)
Women	15,650	9438	63.7 (0.7)		8750	59.2 (0.8)
Age, y				<0.0001			<0.0001
1–3	3747	1376	42.6 (1.3)		1308	40.5 (1.2)
4–8	5162	2108	47.6 (1.2)		1984	44.6 (1.2)
9–13	4897	1439	34.9 (1.1)		1308	31.5 (1.0)
14–18	4505	1187	29.5 (1.2)		1040	26.1 (1.1)
19–30	6088	2167	39.3 (1.0)		1889	34.2 (0.9)
31–50	10,101	4778	51.2 (0.8)		4247	45.4 (0.8)
51–70	10,159	6420	68.7 (0.9)		5826	63.2 (1.0)
>70	4705	3575	80.6 (0.8)		3348	76.1 (0.8)
Race/Hispanic origin3				<0.0001			<0.0001
Non-Hispanic white	18,261	10,225	58.6 (0.8)		9377	53.5 (0.7)
Non-Hispanic black	11,114	4321	40.0 (0.7)		3991	36.8 (0.7)
Hispanic	14,120	5402	38.3 (0.7)		4763	33.3 (0.6)
Family income				<0.0001			<0.0001
Low (PIR <130%)	16,924	5884	37.7 (0.8)		5347	34.2 (0.8)
Middle (PIR 131–350%)	16,497	8073	52.1 (0.8)		7314	47.1 (0.7)
High (PIR >350%)	11,712	7157	63.3 (0.8)		6557	57.9 (0.8)
Household food security				<0.0001			<0.0001
Food insecure	10,617	3691	38.3 (0.9)		3287	33.6 (0.9)
Food secure	38,020	19,005	55.6 (0.6)		17,347	50.8 (0.6)
Health insurance				<0.0001			<0.0001
Private	23,203	12,887	59.0 (0.6)		11,786	53.8 (0.6)
Public	17,389	7075	47.1 (0.8)		6481	43.2 (0.8)
None	8366	2933	38.3 (1.0)		2554	33.3 (0.9)
Screen time (2–18 y)4				<0.0001			<0.0001
≤1 h/d	1783	709	46.8 (2.3)		651	43.4 (2.4)
>1–2 h/d	2614	1054	45.7 (1.5)		981	42.6 (1.6)
>2–4 h/d	4925	1838	41.7 (1.1)		1688	38.1 (1.1)
>4 h/d	5132	1594	33.8 (1.2)		1455	30.4 (1.1)
Educational level (≥19 y)				<0.0001			<0.0001
Less than high school	7358	3244	42.8 (0.7)		2855	37.9 (0.8)
High-school graduate or equivalent	6957	3511	52.4 (1.1)		3182	47.0 (1.0)
Some college/associate degree	8928	5195	59.2 (0.9)		4740	53.9 (0.8)
College degree or greater	6960	4723	68.9 (0.7)		4297	63.1 (0.8)
Weight status (≥19 y)5				0.0001			<0.0001
Underweight	867	458	56.0 (2.2)		414	48.8 (2.0)
Normal	8311	4594	59.3 (0.9)		4241	55.0 (0.9)
Overweight	9957	5504	58.6 (0.8)		4965	53.2 (0.9)
Obese	11,801	6312	55.4 (0.8)		5623	49.4 (0.7)
Smoking status (≥19 y)				<0.0001			<0.0001
Never	17,013	9624	59.4 (0.6)		8733	54.1 (0.6)
Former	7340	4692	66.7 (0.9)		4268	60.9 (1.0)
Current, occasional	1195	506	43.1 (2.3)		436	37.1 (2.1)
Current, daily	5042	1979	41.9 (1.1)		1750	36.6 (1.1)
Alcohol use, drinks/d (≥19 y)6				<0.0001			<0.0001
0	8469	4775	58.0 (1.0)		4368	53.4 (1.0)
1	6852	4425	67.4 (0.9)		4078	62.3 (1.0)
2	5401	3075	60.9 (1.0)		2746	54.7 (1.0)
≥3	7310	3060	44.8 (1.1)		2675	39.1 (1.0)
Self-reported health status (≥19 y)				<0.0001			<0.0001
Excellent or very good	10,285	5999	61.8 (0.8)		5465	56.6 (0.8)
Good	11,921	6378	55.1 (0.8)		5750	49.5 (0.8)
Fair or poor	7136	3695	53.3 (0.9)		3315	48.1 (0.8)

¹Estimates are percentages (SE) adjusted for NHANES interview survey weights, unless otherwise noted. Sample sizes in this table vary by demographic and lifestyle characteristic category either due to missing or unavailable data in certain age groups (e.g., <2 y). Although prevalence estimates reported in this table are nationally representative (i.e., weighted estimates), the sample sizes do not reflect nationally representative estimates, but rather the study population. DS, dietary supplement; DSMQ, Dietary Supplement and Prescription Medicine Questionnaire; MN, micronutrient-containing; PIR, family poverty-to-income ratio; 24HR, 24-h dietary recall.

² P values are from Satterthwaite-adjusted Wald χ² tests.

³Hispanic includes those who identified as Mexican American. The “other” race and Hispanic origin category is not presented herein but is represented in the overall prevalence estimates.

⁴Data on screen time were only available for children aged 2–11 y in NHANES 2007–2010.

⁵Weight status is classified according to BMI category for those (≥19 y) with complete and available data (underweight: <18.5 kg/m²;  normal weight: 18.5–24.9 kg/m²; overweight: 25.0–29.9 kg/m²; obese: ≥30 kg/m²).

⁶An alcoholic drink refers to 12 ounces of beer, 5 ounces of wine, or 1½ ounces of liquor.

These group differences were also observed when evaluating the prevalence of any and MN DS use using the DSMQ alone, as opposed to the DSMQ or at least one 24HR, although the overall prevalence of DS use was lower in the total population and across all population subgroups (Supplemental Table 2). In the total population the prevalence of any DS use estimated from the DSMQ or at least one 24HR was 53%, whereas estimated from the DSMQ alone it was only 48% (Table 2; Supplemental Table 2). Similar findings were apparent when evaluating the prevalence of MN supplement use in the total population; nearly half (48%) of the US population was estimated to take an MN DS on the DSMQ or at least one 24HR (Table 2), as compared with a lower prevalence of MN DS use (43%) when using the DSMQ alone (Supplemental Table 2).

Trends in the prevalence of DS use

DS use in the US population (≥1 y) increased from 49% in 2007 to 56% in 2018 (P < 0.001; Supplemental Figure 1). Use of MN products also increased during this period, from 46% to 50% (P = 0.02). Differences in DS consumption patterns in population subgroups were also noted. Namely, across the study period (2007–2018), the prevalence of any and MN DS use varied by age, with a larger proportion of adults than children, on average, taking a DS. For most age groups, DS usage patterns remained relatively stable over time, with the exception of younger (19–30 y) and older adults (>70 y) (Figure 1). In younger adults (19–30 y) DS use increased from 36% to 41%; and in older adults use increased from 76% in 2007–2010 to 83% in 2015–2018. For older adults, MN use also increased from 71% to 78% over the decade, with a steeper increase in use about halfway through the study period that then stabilized by 2015–2018.

FIGURE 1.

Trends in the prevalence of overall (A) and micronutrient-containing (B) supplement use in the US population, by age group, NHANES 2007–2018. Estimates are percentages adjusted for NHANES interview survey weights, unless otherwise noted. Linear trends were tested using orthogonal polynomial contrasts. *Indicates significant increasing linear trend from 2007–2010 to 2015–2018, P < 0.05. DS, dietary supplement; MN, micronutrient containing.

In adults, noteworthy changes in the prevalence of supplement use by sex, race and Hispanic origin, and family income were observed over the 10-y span. Supplement use increased in both men (48% to 55%; P = 0.002) and women (60% to 66%; P = 0.002), but especially in men, where the prevalence of use increased not only for any DS over time, but also for MN products (42% to 47%; P = 0.01) (Table 3). DS usage patterns also varied by race and Hispanic origin; non-Hispanic whites consistently had a greater prevalence of any and MN DS use than their non-Hispanic black and Hispanic counterparts across the decade. Even so, a larger proportion of non-Hispanic black and Hispanic adults were more likely to take a DS in 2018 than 2007, narrowing the race and Hispanic origin gap in use by the end of the study period (2017–2018). Similarly, adults living in high-income (PIR >350%) households tended to have higher DS use when compared with those in low- and middle-income households; but DS use significantly increased in adults between 2007 and 2018, regardless of household income level. For example, DS use increased from 62% to 68% for high-income, from 53% to 59% for middle-income, and from 38% to 48% for low-income adults during the study period. Differential patterns in DS use in adults by food security status also persisted (43% food insecure compared with 60% food secure); yet the prevalence of any DS and MN DS increased in both the food insecure and the food secure. These findings were especially notable in adults living with food insecurity; the prevalence of DS use in this subpopulation increased from 36% in 2007 to 51% in 2018 (Table 3).

TABLE 3.

Trends in the prevalence of overall and micronutrient-containing supplement use in nonpregnant, nonlactating US adults (>18 y), by sex, race/Hispanic origin, and family income, NHANES 2007–2018¹

	Overall	2007–2008 (n = 5477)	2009–2010 (n = 5820)	2011–2012 (n = 4878)	2013–2014 (n = 5076)	2015–2016 (n = 5036)	2017–2018 (n = 4777)	P for trend2
	Reported on DSMQ or at least one 24HR
Among all adults (≤19 y)
Any DS	57.6 (0.6)	53.9 (2.0)	54.5 (1.1)	57.6 (1.1)	58.9 (1.5)	59.5 (1.7)	60.6 (1.5)	0.0006
MN DS	52.2 (0.6)	49.5 (2.1)	49.3 (1.0)	52.8 (1.2)	54.2 (1.4)	53.1 (1.5)	54.0 (1.4)	0.0120
Sex
Men
Any DS	51.2 (0.7)	47.6 (2.3)	48.9 (1.7)	50.9 (1.4)	51.3 (1.7)	53.3 (1.7)	54.7 (1.3)	0.0019
MN DS	44.9 (0.7)	42.0 (2.4)	41.9 (1.4)	45.1 (1.3)	46.6 (1.4)	46.2 (1.6)	47.3 (1.4)	0.0094
Women
Any DS	63.7 (0.8)	60.0 (2.0)	59.9 (1.1)	64.2 (1.4)	66.3 (2.0)	65.5 (2.0)	66.3 (2.1)	0.0025
MN DS	59.2 (0.8)	56.7 (2.0)	56.3 (1.2)	60.4 (1.8)	61.7 (1.9)	59.7 (2.2)	60.5 (2.1)	0.0688
Race/Hispanic origin
Non-Hispanic white
Any DS	62.3 (0.8)	59.0 (2.9)	58.8 (1.4)	63.5 (1.4)	64.9 (1.5)	64.3 (2.0)	63.6 (1.9)	0.0322
MN DS	56.9 (0.8)	54.5 (2.9)	53.4 (1.2)	58.5 (1.7)	60.4 (1.4)	57.4 (1.8)	57.1 (1.9)	0.1540
Non-Hispanic black
Any DS	46.2 (0.8)	42.1 (2.0)	44.3 (1.6)	47.1 (1.8)	44.8 (1.8)	48.2 (2.6)	50.5 (1.4)	0.0013
MN DS	42.3 (0.7)	39.1 (2.0)	40.4 (1.4)	43.3 (1.9)	41.2 (1.5)	43.8 (2.3)	45.5 (1.5)	0.0089
Hispanic
Any DS	43.8 (0.8)	39.2 (2.6)	40.3 (1.7)	39.0 (1.5)	43.6 (1.3)	45.0 (1.7)	53.8 (2.1)	<0.0001
MN DS	37.6 (0.7)	33.3 (2.5)	35.6 (1.5)	34.0 (1.3)	37.1 (1.3)	38.8 (1.4)	45.2 (2.0)	0.0001
Family income
PIR <130%
Any DS	43.1 (0.8)	38.0 (2.0)	39.3 (1.6)	43.5 (1.7)	46.1 (2.3)	43.5 (1.8)	47.6 (2.5)	0.0006
MN DS	39.0 (0.8)	34.4 (1.6)	35.2 (1.7)	38.6 (1.7)	42.3 (2.4)	40.5 (1.8)	42.7 (2.4)	0.0004
PIR 131–350%
Any DS	56.3 (0.9)	53.4 (3.1)	53.0 (2.5)	55.7 (1.4)	58.4 (1.9)	57.6 (1.8)	59.5 (2.5)	0.0351
MN DS	50.7 (0.9)	48.8 (2.9)	47.4 (2.5)	50.8 (1.3)	53.1 (1.7)	51.0 (1.9)	53.0 (2.7)	0.1227
PIR >350%
Any DS	66.4 (0.8)	62.1 (2.4)	63.6 (1.2)	68.7 (1.7)	68.2 (1.5)	68.2 (2.6)	67.7 (2.1)	0.0265
MN DS	60.6 (0.9)	57.5 (2.8)	58.0 (1.6)	64.0 (1.8)	63.3 (1.7)	60.2 (2.4)	60.5 (1.8)	0.2799
Food security
Food insecure
Any DS	43.1 (1.0)	35.7 (1.9)	38.8 (2.2)	42.4 (2.5)	43.2 (2.4)	42.9 (2.4)	51.4 (3.4)	0.0001
MN DS	37.7 (1.0)	32.4 (1.8)	33.3 (2.1)	36.6 (2.6)	38.6 (2.2)	37.3 (2.3)	44.4 (3.4)	0.0010
Food secure
Any DS	60.1 (0.6)	56.1 (2.1)	57.0 (1.2)	60.4 (1.1)	61.8 (1.6)	63.1 (1.6)	62.6 (1.4)	0.0004
MN DS	54.7 (0.6)	51.5 (2.2)	51.8 (1.2)	55.8 (1.3)	57.0 (1.5)	56.4 (1.4)	56.1 (1.3)	0.0082

¹Estimates are percentages (SE) adjusted for NHANES interview survey weights, unless otherwise noted. DS, dietary supplement; DSMQ, Dietary Supplement and Prescription Medicine Questionnaire; MN, micronutrient containing; PIR, family poverty-to-income ratio; 24HR, 24-h dietary recall.

²Linear trends were tested using orthogonal polynomial contrasts; significance was set at P < 0.05.

In children (1–18 y), any DS and MN DS use remained relatively stable over time, with ∼38% reporting any DS use between 2007–2008 and 2017–2018 (Table 4). These findings were consistent across most population subgroups, including those defined by sex, race and Hispanic origin, and family income. Across the study period, the prevalence of DS use in girls and boys was comparable, with 38% of girls and 37% of boys using a DS; although a slight increase in any DS use was observed in girls, from 35% in 2007 to 42% in 2018 (P = 0.04). Non-Hispanic white children (44%) were also more likely to take a DS when compared with non-Hispanic black (24%) and Hispanic children (27%), which remained constant over the decade. When evaluating DS use in children from 2007–2008 through 2017–2018 by socioeconomic indicators, supplement use increased with household income and food security, but remained stable over time (Table 4). The exception was children living with food insecurity, whose DS use increased from 22% in 2007 to 31% in 2018 (P = 0.005), narrowing the gap in DS use between food-secure and food-insecure children.

TABLE 4.

Trends in the prevalence of overall and micronutrient-containing supplement use in US children (1–18 y), by sex, race/Hispanic origin, and family income, NHANES 2007–2018¹

	Overall	2007–2008 (n = 3233)	2009–2010 (n = 3403)	2011–2012 (n = 3183)	2013–2014 (n = 3093)	2015–2016 (n = 2978)	2017–2018 (n = 2427)	P for trend2
	Reported on DSMQ or at least one 24HR
Among all children (1–18 y)
Any DS	37.7 (0.8)	36.0 (2.2)	37.7 (1.3)	38.2 (1.8)	38.5 (2.0)	37.1 (2.7)	38.9 (1.9)	0.4506
MN DS	34.6 (0.8)	34.9 (2.1)	35.2 (1.1)	35.3 (1.8)	35.4 (1.7)	33.2 (2.5)	33.6 (1.9)	0.4633
Sex
Boys
Any DS	37.1 (1.0)	36.5 (2.3)	38.4 (1.4)	39.5 (2.4)	36.3 (2.3)	36.1 (2.9)	35.5 (3.0)	0.4894
MN DS	34.0 (0.9)	35.4 (2.1)	35.7 (1.3)	36.1 (2.3)	33.2 (2.2)	31.9 (2.8)	30.7 (2.8)	0.0607
Girls
Any DS	38.3 (1.0)	35.5 (2.7)	37.0 (1.5)	37.0 (1.7)	40.7 (2.4)	38.1 (3.1)	42.5 (2.2)	0.0452
MN DS	35.3 (0.9)	34.2 (2.7)	34.7 (1.4)	34.4 (1.9)	37.6 (1.9)	34.6 (2.8)	36.7 (2.2)	0.4548
Race/Hispanic origin
Non-Hispanic white
Any DS	44.2 (1.2)	42.8 (3.9)	43.9 (1.4)	44.2 (2.7)	44.7 (3.0)	45.7 (2.5)	44.3 (2.8)	0.6021
MN DS	40.6 (1.1)	41.4 (3.6)	40.8 (1.4)	40.1 (3.0)	41.1 (2.7)	40.8 (2.5)	38.8 (2.9)	0.6327
Non-Hispanic black
Any DS	23.9 (1.0)	23.3 (1.9)	22.9 (1.8)	25.5 (2.0)	26.4 (3.6)	22.6 (2.4)	23.0 (2.7)	0.9353
MN DS	22.7 (0.9)	22.7 (1.9)	21.0 (2.0)	24.6 (2.0)	25.0 (3.3)	21.6 (2.5)	21.0 (2.5)	0.7292
Hispanic
Any DS	27.4 (1.0)	24.8 (1.3)	26.6 (2.7)	27.8 (1.8)	28.1 (2.1)	23.0 (3.0)	34.2 (2.8)	0.0682
MN DS	24.8 (1.0)	23.7 (1.5)	24.6 (2.4)	26.1 (1.8)	25.1 (2.2)	20.6 (2.9)	28.8 (2.5)	0.5086
Family income
PIR <130%
Any DS	26.6 (1.1)	25.0 (2.6)	26.0 (2.2)	26.2 (2.7)	28.3 (2.4)	27.2 (3.1)	27.5 (3.3)	0.4493
MN DS	24.1 (1.0)	23.9 (2.5)	23.7 (1.6)	24.6 (2.2)	25.8 (2.6)	23.1 (3.0)	23.0 (2.4)	0.8093
PIR 131–350%
Any DS	39.6 (1.1)	37.3 (2.1)	41.5 (2.4)	41.2 (2.9)	41.4 (2.0)	36.4 (2.8)	40.2 (3.3)	0.9734
MN DS	36.2 (1.0)	35.8 (1.9)	38.9 (2.3)	35.6 (3.1)	38.8 (1.6)	33.9 (2.7)	34.4 (3.1)	0.3800
PIR >350%
Any DS	49.0 (1.4)	47.4 (3.3)	46.1 (3.1)	50.9 (3.0)	50.8 (4.4)	48.8 (3.2)	51.3 (4.1)	0.3564
MN DS	45.7 (1.4)	46.2 (3.4)	43.7 (2.6)	49.0 (3.7)	46.1 (4.5)	43.2 (2.5)	45.8 (4.1)	0.8238
Food security
Food insecure
Any DS	26.6 (1.2)	22.3 (2.4)	20.4 (1.5)	28.5 (3.5)	27.9 (2.9)	27.4 (3.2)	31.4 (3.2)	0.0053
MN DS	23.5 (1.2)	20.9 (2.2)	19.2 (1.4)	24.1 (2.7)	24.4 (2.1)	23.1 (3.8)	28.5 (3.0)	0.0292
Food secure
Any DS	40.5 (0.9)	38.7 (2.4)	41.1 (1.4)	40.7 (1.9)	41.1 (2.3)	40.3 (2.5)	41.4 (2.5)	0.5639
MN DS	37.4 (0.9)	37.6 (2.3)	38.5 (1.2)	38.1 (2.1)	38.1 (2.2)	36.5 (2.3)	35.3 (2.5)	0.3578

¹Estimates are percentages (SE) adjusted for NHANES interview survey weights, unless otherwise noted. DS, dietary supplement; DSMQ, Dietary Supplement and Prescription Medicine Questionnaire; MN, micronutrient containing; PIR, family poverty-to-income ratio; 24HR, 24-h dietary recall.

²Linear trends were tested using orthogonal polynomial contrasts; significance was set at P < 0.05.

Types of DS products used

MVM DSs (63%) remain the most common type of DS taken by DS users (≥1 y), but their use has significantly decreased over time (P < 0.0001; Table 5). Other commonly consumed MN products include calcium and vitamin D (11%), MVs (8%), MMs (3%), and single-nutrient DSs, such as calcium (7%), iron (5%), magnesium (4%), and vitamins B-12 (8%), C (14%), D (20%), and E (5%). Whereas MV and MM use remained stable over the 10-y span, use of calcium and vitamin D products decreased, and some single-nutrient DSs, including magnesium, vitamin B-12, and vitamin D, showed significant increases in use (all P < 0.001). Specifically, use of magnesium products increased from 2% in 2007 to 8% in 2018, and use of vitamin B-12 and vitamin D increased from 5% to 12% and 5% to 30%, respectively, during the same period. In contrast, use of some MN products decreased between 2007 and 2018, namely products containing vitamins C and E. Use of most non-MN DS products was low (i.e., <5%) in the population across the study period, with the exception of ω-3 fatty acids (22%), botanical and herbal (12%), and joint (8%) supplements (Table 5).

TABLE 5.

Trends in the prevalence of select micronutrient-containing and non–micronutrient-containing supplement product types in the US population (≥1 y) who report taking any dietary supplement, NHANES 2007–2018¹

	Overall	2007–2008 (n = 3791)	2009–2010 (n = 4149)	2011–2012 (n = 3753)	2013–2014 (n = 3884)	2015–2016 (n = 3787)	2017–2018 (n = 3686)	P for trend2
	Reported on DSMQ or at least one 24HR
MN DS
MVM	62.6 (0.6)	69.6 (1.2)	68.3 (1.7)	61.8 (1.2)	65.0 (1.2)	57.5 (1.3)	56.4 (1.5)	<0.0001
MV	8.2 (0.3)	9.1 (0.8)	6.7 (0.7)	8.6 (0.7)	7.2 (0.6)	8.7 (0.6)	8.8 (0.9)	0.6430
MM	2.9 (0.2)	2.9 (0.5)	2.6 (0.3)	3.1 (0.6)	2.4 (0.4)	3.2 (1.0)	3.2 (0.4)	0.5702
Calcium	7.3 (0.3)	8.0 (1.0)	6.6 (0.4)	6.8 (0.6)	9.2 (0.7)	7.0 (0.7)	6.3 (0.7)	0.4774
Calcium and vitamin D	11.3 (0.3)	13.8 (0.8)	14.3 (0.7)	13.4 (0.9)	10.6 (0.8)	8.8 (0.9)	7.7 (0.9)	<0.0001
Iron	4.5 (0.2)	3.7 (0.3)	4.1 (0.3)	4.2 (0.5)	4.5 (0.5)	4.6 (0.4)	5.7 (0.7)	0.0083
Magnesium	4.0 (0.3)	1.7 (0.3)	2.0 (0.3)	3.2 (0.7)	3.8 (0.5)	5.5 (0.6)	7.6 (0.9)	<0.0001
Potassium	2.4 (0.2)	2.2 (0.4)	1.8 (0.3)	1.2 (0.2)	2.2 (0.4)	3.8 (0.6)	3.3 (0.4)	0.0008
Zinc	1.4 (0.1)	1.8 (0.2)	1.4 (0.3)	1.2 (0.3)	1.8 (0.4)	1.0 (0.2)	0.8 (0.2)	0.0073
Vitamin A	0.8 (0.1)	0.6 (0.1)	1.0 (0.2)	1.2 (0.3)	0.9 (0.3)3	0.5 (0.1)	0.4 (0.1)	0.0274
Niacin	1.1 (0.1)	1.6 (0.4)	1.2 (0.2)	1.6 (0.3)	0.9 (0.3)3	1.1 (0.2)	0.6 (0.2)3	0.0100
Vitamin B-6	1.2 (0.1)	1.5 (0.3)	1.4 (0.3)	1.1 (0.3)	1.2 (0.3)	0.9 (0.3)3	1.3 (0.4)3	0.4558
Folic acid	2.3 (0.1)	2.6 (0.4)	2.6 (0.2)	2.0 (0.3)	2.3 (0.3)	2.2 (0.4)	2.3 (0.4)	0.4423
Vitamin B-12	8.2 (0.3)	5.3 (0.6)	5.6 (0.3)	8.0 (0.8)	8.3 (0.6)	9.4 (1.1)	12.1 (0.8)	<0.0001
Vitamin C	13.7 (0.4)	15.2 (1.2)	13.2 (0.9)	15.2 (0.8)	13.9 (0.6)	13.0 (1.3)	11.7 (0.7)	0.0238
Vitamin D	19.6 (0.6)	5.0 (0.5)	12.7 (1.2)	20.8 (1.5)	22.6 (1.1)	24.9 (1.3)	29.7 (1.5)	<0.0001
Vitamin E	4.9 (0.2)	8.0 (0.5)	5.0 (0.5)	5.6 (0.8)	4.4 (0.5)	3.6 (0.3)	3.0 (0.3)	<0.0001
Non-MN DS
Omega-3 fatty acids/fish oil	22.1 (0.5)	19.6 (1.4)	22.1 (1.2)	24.6 (1.3)	23.2 (0.8)	22.4 (1.5)	20.4 (1.2)	0.7659
Botanicals/herbals	12.0 (0.5)	11.3 (0.9)	10.6 (0.8)	12.1 (1.0)	11.5 (1.1)	13.9 (1.6)	12.4 (1.0)	0.0985
Probiotics	4.4 (0.4)	1.1 (0.2)	1.4 (0.3)	3.0 (0.4)	3.6 (0.5)	8.7 (1.3)	7.7 (1.2)	<0.0001
Fiber	4.2 (0.2)	5.0 (0.4)	3.8 (0.5)	4.5 (0.7)	4.5 (0.4)	4.2 (0.4)	3.6 (0.4)	0.0687
Coenzyme Q10	3.3 (0.3)	2.4 (0.5)	2.0 (0.4)	3.2 (0.6)	3.2 (0.4)	4.1 (0.9)	4.9 (0.7)	0.0006
Joint supplements	7.6 (0.3)	9.3 (0.6)	7.8 (0.6)	6.4 (0.6)	8.3 (0.8)	7.4 (1.0)	6.3 (0.6)	0.0129
Protein/sports	2.6 (0.2)	2.2 (0.4)	1.5 (0.2)	1.9 (0.3)	2.6 (0.6)	2.7 (0.4)	4.6 (0.4)	<0.0001
Echinacea	0.7 (0.1)	1.3 (0.3)	ES4	0.7 (0.2)	0.7 (0.1)	ES4	0.7 (0.2)3	—
Melatonin	2.1 (0.2)	0.4 (0.1)	0.7 (0.2)	1.6 (0.5)3	2.1 (0.5)	3.3 (0.6)	4.3 (0.5)	<0.0001

¹Estimates are percentages (SE) adjusted for NHANES interview survey weights, unless otherwise noted. DS, dietary supplement; DSMQ, Dietary Supplement and Prescription Medicine Questionnaire; MM, multimineral; MN, micronutrient containing; MV, multivitamin; MVM, multivitamin-mineral; NCHS, National Center for Health Statistics; 24HR, 24-hour dietary recall.

²Linear trends were tested using orthogonal polynomial contrasts; significance was set at P < 0.05. The dash (—) indicates a suppressed P for trend value due to suppressed estimates and/or nonapplicable estimates.

³The relative SE is >30% but ≤40% and might be statistically unreliable.

⁴ES indicates that the estimate is suppressed due to a statistically unreliable SE (i.e., relative SE ≥40%), per NCHS analytical guidelines.

In adult DS users, patterns of use for specific product types generally paralleled those observed in the total population (Supplemental Table 3). MVM use remained high (60%) despite decreasing over time (P < 0.0001), whereas MV (9%) and MM (3%) use remained stable. Calcium and vitamin D products were also prevalent (13%), but their use decreased across the decade (P < 0.0001). For single-nutrient products, >5% of US adults used supplements containing calcium, iron, magnesium, or vitamins B-12, C, D, or E. Whereas use of supplements containing iron, vitamin B-12, or vitamin D increased over the 10-y span (all P < 0.02), use of calcium did not change, and the use of vitamin C and E products decreased (both P < 0.005). ω-3 fatty acids (26%) and botanical and herbal (14%) DSs were the most commonly consumed non-MN products in adults, and their use did not vary across the study period.

Similar to adults, MVM use by child DS users was very high (77%) overall, but decreased from 82% in 2007 to 70% in 2018 (Supplemental Table 4). However, unlike MVM use, MV use in children differed from that of adults and decreased from 8% to 3% over the decade (P < 0.0001). Use of calcium and vitamin D (3%) and MM (<1%) products was low in children, along with the prevalence of single-vitamin or -mineral supplements that contained the following nutrients (all <2%):  calcium, folic acid, iron, magnesium, zinc, and vitamins A, B-12, and E (Supplemental Tables 1 and 4). Use of non-MN products (e.g., fiber, melatonin, probiotics, and botanical/herbal) was also minimal in children, apart from the use of ω-3 fatty acid (6%) supplements.

Mean nutrient intake from supplements

The estimated mean daily intakes of micronutrients from DSs by users (≥1 y) who took a DS containing the specified nutrients of interest are presented in Supplemental Table 5. Trends in the amounts of nutrient intake from supplemental sources varied by nutrient, with notable changes in intake over time for calcium, iron, vitamin B-12, and vitamin D. Calcium intake from supplemental sources decreased from 360 mg/d in 2007 to 276 mg/d in 2018, whereas intake of iron (17 mg/d to 20 mg/d), vitamin B-12 (76 mg/d to 235 mg/d), and vitamin D (12.2 μg/d to 38 μg/d) increased during the same period. Moreover, DS intake of other micronutrients, such as folic acid, vitamin B-6, and vitamin C, remained stable, at ∼355 mg/d, ∼8 mg/d, and ∼196 mg/d, respectively.

Discussion

Between 2007–2008 and 2017–2018, the prevalence of DS use increased in the US population, not only in terms of any supplement used but also in the number of MN product types used. These findings can be attributable to increases in the use of single-nutrient products, such as magnesium and vitamins B-12 and D, as opposed to MVMs, which declined in use over the decade. However, DS usage patterns differed by population subgroup across the study period. Most notably, the increase in overall DS use was largely driven by increases in use among adults, rather than children, whose DS use remained stable from 2007 to 2018.

These findings are consistent with previous literature that suggests DS use in children has remained unchanged and is largely attributed to the use of MN (e.g., MVM) products, regardless of population subgroup (11, 27, 28). However, differences in the prevalence of DS use in children over the study period were most marked for girls when compared with boys, and for children living with food insecurity when compared with their food-secure peers. Overall supplement use among girls increased 20% from 2007 to 2018 (P = 0.045), whereas use among boys remained stable over the same period. Qato et al. (28) found that significant sex differences in DS use exist among adolescents, with adolescent girls having a higher prevalence of any DS use than boys. Adolescent girls are also more likely to consume MN products, whereas adolescent boys predominantly consume non-MN products (e.g., ω-3 fatty acids, protein) (28). Adolescence is a critical period of transition in terms of dietary patterns, as well as growth acceleration and sexual maturation, and therefore meeting nutrient needs in this population is important (56). Adolescents tend to have the poorest diet quality of any age group (57) and have the lowest prevalence of DS use among all age groups. A recent study found that adolescents (14–18 y), regardless of sex, were at high risk of micronutrient inadequacy for some micronutrients, including vitamins A (food sources only), C, D, and E (food sources only), as well as calcium, magnesium, and choline (58). Moreover, adolescent girls face unique nutrient needs due to menstruation and require additional iron to offset menstrual losses as a result (59); ∼16.4% of food-secure and ∼18.6% of food-insecure adolescent girls are at risk of inadequate iron intake (58), and ∼12–13% are considered iron deficient based on measures of total body iron (58). Thus, poor iron nutriture in adolescent girls, one of the few clinical indications of inadequate nutrient intake in the United States, could be improved by increasing intakes of iron-rich foods, such as lean meats and iron-fortified cereals, and by targeted use of iron supplements (58).

The increase in DS use in children living with food insecurity is an important finding, given that food insecurity among children is associated with compromised nutrient intake for several micronutrients and is associated with lower DS use (60). Jun et al. (58) found that differences in vitamin D intake among food-secure and food-insecure children were substantial, especially when including intake from DSs. Vitamin D is a nutrient of public health concern (61), and adequate vitamin D intake is especially important for peak bone mass development in children (62). Therefore, increased intake of dietary sources of vitamin D, such as fatty fish and vitamin D–fortified foods, and targeted use of DSs containing vitamin D could be helpful for improving vitamin D intakes in children living with food insecurity. The risk of inadequate vitamin D intake has been estimated to be 81.0% in food-insecure boys and 92.8% in food-insecure girls, even with the inclusion of DSs (58). The 2020–2025 Dietary Guidelines for Americans recommend that nutrient needs be primarily met through the intake of foods and beverages, but recognize that for certain population subgroups, like adolescent girls, supplementation of some nutrients (e.g., iron) can be warranted (61).

Supplement use in adults increased between 2007 and 2018 across most population subgroups, including those defined by sex, race and Hispanic origin, family income, and household food security status. These increases in overall use were primarily driven by increases in the prevalence of use in younger (19–30 y) and older (>70 y) adults, as opposed to middle-aged adults (31–50 y; 51–70 y), whose DS use remained stable across the decade. An earlier study investigating trends in DS use in adults also reported that DS use had stabilized in middle-aged adults, while increasing in older adults (7). DS use in older adults is higher than any other age group, and many older adults take multiple DS products (8). Although the DRI reports recommend that older adults supplement their diet with DSs to increase their likelihood of meeting nutrient needs for vitamins B-12 and D (63), most older adults report other health-related primary motivations for their DS use, such as to improve or maintain health (8, 44). MN products have the potential to help mitigate nutrient inadequacy in older adults, but they also can increase the risk of nutrient excess for some nutrients (e.g., those also present in fortified foods) (17). Thus, caution is warranted when multiple DSs are consumed daily.

US consumers spent upwards of 50 billion on supplements in recent years (10), but their use was predominantly driven by personal choice, as opposed to recommendations by a health care practitioner. Previous studies have reported that only 18% of children and 23% of adults in the United States took a DS at the recommendation of a practitioner (6, 44).

Our previous work (1, 5, 6, 27, 44, 64) and the work of others (7, 11, 12, 28, 65) found that approximately one-third of children and one-half of adults in the United States take a DS when data from the DSMQ were used alone. The prevalence of supplement use in children (∼38%) and adults (∼58%) estimated in the present study was slightly higher than previous reports because we used both the DSMQ and the 24HRs, which leverages the strengths of both instruments and increases the length of the reference time period. Moreover, a study evaluating DS assessment methods in adults found that >500 participants (i.e., 8.8% of DS users) in NHANES 2011–2014 reported DS products in the 24HR that they had not reported on the DSMQ (24). Therefore, the use of a combined approach (i.e., the DSMQ or at least one 24HR method) can potentially capture a larger fraction of DS products consumed, in addition to identifying a larger proportion of DS users (24). The same study also found that when only the DSMQ was used to estimate DS use in adults the prevalence was 52%, and a higher (57%) prevalence of use was estimated when both the DSMQ and the 24HRs were used (24). Similarly, Nicastro et al. (25) determined that the estimated prevalence of MVM use also differed by the DS instrument of choice, and that MVM prevalence estimates were higher when using data from both the DSMQ and the 24HR. For example, 29.3% of men and 35.5% of women reported MVM use on the DSMQ, 26.3% of men and 33.2% of women reported use on the 24HR, and ∼32% and ∼40% reported MVM use, respectively, when using the DSMQ and 24HR combined (25). These findings are congruent with those of the present study, which found that an estimated 53% of the US population took a DS when both the DSMQ and the 24HRs were used, but only 48% took a DS when only the DSMQ was used. In summary, this study contributes to a small body of literature that uses multiple methods of DS assessment designed to improve the comprehensive assessment of DS exposure.

Limitations and strengths

A number of caveats and considerations are worth noting for the present analysis. Very little is known about the reporting errors that exist for measuring DS intake and use. In children, reporting errors can arise if DS use (i.e., frequency and dose) information is reported by nonparent proxies. Moreover, DS product formulations are continuously changing and could have been reformulated over the data collection period; but NHANES staff strive to ensure that current DS formulations at the time of data collection are matched to the DS reported. DS users can use multiple DS products daily, which can create very large and extreme values that can exert undue influence on estimates of mean intakes of nutrients from DS.

Classification of DS products into different product categories can also present several challenges. Firstly, some products can be classed into multiple categories, meaning that subjective decisions are required to create mutually exclusive product categories for analysis at the product level because no standardized definitions for categories exist. Secondly, DS formulations that primarily contain a single vitamin or mineral (e.g., calcium) often contain trace amounts of other nutrients for technical function (e.g., binders, fillers, colors, excipients). In such cases, NHANES considers these products to contain >1 vitamin and/or mineral; accordingly, decisions on the best approach for classifying these products are required. Lastly, in addition to trace vitamins or minerals, MN DSs, such as MVM, MV, MM, calcium and vitamin D, and single-nutrient products, also can contain herbals and botanicals or other ingredients that can affect the ability to classify them as an MN DS, especially with a single vitamin or mineral DS classification.

NHANES relies on label declarations when evaluating DS use. Analytically derived values for most micronutrients tend to be higher than labeled amounts for MVMs, but little is known about other product types and single-nutrient DSs. In NHANES, participants are also probed to recall DSs previously reported on the DSMQ during the 24HRs, but this is not a common practice on 24HRs in other settings. Future work improving our understanding of DS assessment and associated measurement error is warranted (17).

Given the cross-sectional nature of this analysis, we are unable to draw inferences regarding the reasons for changes in DS use over time that occur at the individual level. Dietary choices, food sources of nutrients, and lifestyle preferences shift over time, and these shifts can in turn alter the types of DS used, the motivations for their use, and the nutritional status of the US population (41).

Finally, the present analysis was restricted to the US population (≥1 y) and did not include those aged <1 y or pregnant and/or lactating women due to inadequate sample sizes when stratifying by NHANES cycle. However, Gahche et al. (66) reported that 18.2% of infants and toddlers took ≥1 DS/d in NHANES 2007–2014, and Jun et al. (67) found that DS use in pregnant (77.4%) and lactating women (70.3%) remained stable from NHANES 1999 to 2014; the NHANES cycles used for these analyses overlap with our study period.

Conclusions

Overall and MN DS use in the US population (≥1 y) increased over the study period, and this can be attributable to increases in the use of single-nutrient products. MVMs remain the most commonly consumed DSs, but their use has decreased over time. DS usage patterns and trends in the prevalence of use varied between population subgroups, including those defined by sex, age, race and Hispanic origin, family income, and food security status.

Notes

This research was funded by a National Institutes of Health/National Cancer Institute grant (NIH/NCI U01CA215834).

Author disclosures: Unrelated to this study, RLB has served as a consultant in the past to the NIH Office of Dietary Supplements, Nestlé, the General Mills Bell Institute, RTI International, and Nutrition Impact; and is a trustee of the International Food Information Council and a former board member of International Life Sciences Institute-North America. RLB has received travel support to present her research on dietary supplements from the Council of Responsible Nutrition, American Society of Nutrition, and the New York Academy of Sciences. She is a member of the Journal's Editorial Board. JTD owns stock in several food and drug companies. Not related to this publication she serves on the scientific advisory committee of the McCormick Science Institute, the Mushroom Council, and Bay State Milling, the Singapore Institute of Food and Biotechnology Innovation (SIFB), and was a one-time consultant for Nestlé in 2020. She serves as an advisor to IAFANS low calorie sweetener committee and on the governance committee of ILSI US-Canada research program and is editor of Nutrition Today. All other authors report no conflicts of interest.

Supplemental Tables 1–5 and Supplemental Figure 1 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn.

Abbreviations used: DS, dietary supplement; DSMQ, Dietary Supplement and Prescription Medicine Questionnaire; MEC, mobile examination center; MM, multimineral; MN, micronutrient containing; MV, multivitamin; MVM, multivitamin-mineral; NCHS, National Center for Health Statistics; PIR, poverty-to-income ratio; 24HR, 24-h dietary recall.

Data Availability

All data utilized in the present manuscript are freely available to the public at: https://wwwn.cdc.gov/nchs/nhanes/default.aspx.