USE OF VITAMIN AND MINERAL SUPPLEMENTS: DEMOGRAPHICS AND AMOUNTS OF NUTRIENTS CONSUMED

Transcription

1 AMEWCAN JOURNAL OF EPIDEMIOLOGY Copyright 199 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol 132, No 6 Printed m U S A USE OF VITAMIN AND MINERAL SUPPLEMENTS: DEMOGRAPHICS AND AMOUNTS OF NUTRIENTS CONSUMED THE 1987 HEALTH INTERVIEW SURVEY AMY F SUBAR AND GLADYS BLOCK Subar, A. F. (NIH, Bethesda, MD 2892), and G. Block. Use of vitamin and mineral supplements: demographics and amounts of nutrients consumed. The 1987 National Hearth Interview Survey. Am J Epidemiol 199;132: Data from the 1987 National Health Interview Survey show that 51.1% of the adults aged years in the United States consumed a vitamin/mineral supplement in the past year, but that only 23.1% did so daily. Whites, women, and older individuals were more likely than blacks, men, and younger individuals to consume supplements regularly. Multivitamins were the most commonly consumed supplement, followed by vitamin C, calcium, vitamin E, and vitamin A. Results suggest that supplementation practices have changed little since the 197s. Results regarding the amounts of nutrients obtained from supplements show that a food frequency type of methodology collects reasonably accurate data reflecting intake of supplements over the past year. Few, if any, individuals were consuming nutrients in amounts considered toxic. Although vitamin and mineral supplementation is a common health habit, it appears not to pose a significant health risk for most of the population. diet; minerals; nutrition surveys; vitamins There is widespread interest in the role of nutrition in disease prevention. Evidence exists that vitamins A, C, and E, and selenium may help to prevent various cancers (1-3). While knowledge of food intake is essential in research regarding the relation of micronutrients to disease, knowledge of supplement use is equally important because use can represent a substantial contribution to total nutrient intake. Further- Received for publication January 24, 1989, and in final form June 28,199. Abbreviations: NHANES I, National Health and Nutrition Examination Survey I; NHANES II, National Health and Nutrition Examination Survey II; NHIS, National Health Interview Survey. From the National Institutes of Health, National Cancer Institute, Division of Cancer Prevention and Control, Bethesda, MD. Reprint requests to Dr. Amy F. Subar, National Institutes of Health, Nationai Cancer Institute, Division of Cancer Prevention and Control, 9 Rockville Pike, EPN 313, Bethesda, MD more, monitoring supplement intake is also important because of the possibility that some persons will ingest nutrients at a toxic level. Several studies have examined supplement use among adults in nonrepresentative samples (4-15), while four studies have done so in nationally representative samples (16-19). Demographic and personal characteristics of supplement users have been examined in detail in two of these national studies (17,19). In additional, several studies have attempted to estimate quantitatively the amount of nutrient intake from supplements (4, 5,1, 14-16, 18-19). Many of these studies, however, have relied upon detailed recording of supplements consumed within a recent brief time period (e.g., within 24 hours prior to interview or over a 3-day period). The purpose of the analysis reported here is to provide estimates of supplement use in a large rep- 191

2 192 SUBAR AND BLOCK resentative sample of the US population, in which supplement use averaged over the entire past year is considered. In addition, this analysis demonstrates the usefulness of a less resource-intensive method of data collection for assessing a population's nutrient intake from supplements. MATERIALS AND METHODS The National Center for Health Statistics and the National Cancer Institute collaborated to conduct the 1987 National Health Interview Survey (NHIS) which included an epidemiology supplement specifically addressing variables of interest to cancer control research. NHIS is a national probability sample survey conducted by household interview (2). Data are presented here from 22,8 adults, aged years, of all races. Subjects answered questions regarding frequency of intake of any supplement and of the following specific supplements known from previous research (12, 17) to be the most commonly consumed: multivitamin/mineral, vitamin A, vitamin C, vitamin E, and calcium. Subjects were asked 1) if there was any consumption in the past year; 2) how many months of the year and days of the month each supplement was usually consumed; 3) how many pills per day were usually consumed; 4) how many milligrams or International Units were in each pill (but subjects were not asked for multivitamins); and 5) if they had consumed the supplement in the past month. For multivitamins, subjects were additionally asked the type of pill usually consumed ("one-a-day," "stress," "therapeutic," or "other"). Because the definition of a supplement "user" varies between studies, we developed six possible user definitions: 1) anytime in the past year, 2) in the past month for >2 days; 3) in the past month for >2 days; 4) every month; 5) every month for >13 days; and 6) every month daily. With these six definitions of user, we could both compare our results with other research having similar definitions and show how varying the definitions affects estimates of the rates of supplement use. In addition to estimates of the proportion who consumed supplements, quantitative estimates of average daily intake of nutrients from supplemental vitamin A, thiamin, riboflavin, vitamin C, vitamin D, vitamin E, iron, and calcium were calculated. The average daily intake of nutrients from supplements was estimated by performing the following calculation: months per year X days per month X pills per day X dosage per pill/365 days. Individuals with invalid data for either months per year or days per month were dropped from the analysis for that particular supplement use calculation. Estimated nutrient amounts for thiamin, riboflavin, vitamin D, and iron were based upon dosages found in multivitamins only and could not take into account possible intake from single nutrient supplement use since such data were not collected. For calcium, the estimate was based on its use as a single nutrient supplement only and does not account for possible use in multivitamins. For vitamins A, C, and E, the estimate was based upon usage from either multivitamins or single nutrient supplements or both. Nutrient content from multiple vitamins was assigned as follows for vitamin A, thiamin, riboflavin, vitamin C, vitamin D, vitamin E, and iron, respectively: one-a-day: 5, IU, mg, mg, 6 mg, 4 IU, 3 IU, and 18 mg; stress: IU, 15 mg, 15 mg, 6 mg, IU, 3 IU, and 15 mg; and therapeutic: 1, IU, 1.3 mg, 1.3 mg, 2 mg, 4 IU, 15 IU, and 12 mg. If subjects were unaware of the type of multivitamin supplement they were consuming (n = 986), the one-a-day-type values were assigned. Subjects were asked to report the dosage usually found in each single supplement of vitamins A, C, and E, and calcium that they consumed. Considerable editing was required, as 1 percent double-keying of the raw data onto the data tape was not per-

3 VITAMIN SUPPLEMENT USE IN THE UNITED STATES 193 formed, and obvious column shifts and other keying errors were detected. Amounts in a single pill were determined to be invalid if they fell outside the range of formulations available on the market: vitamin A, 1 25, IU; vitamin C, 5-1, mg; vitamin E, 5-1, IU; and calcium, 5-8 mg. Individuals with dosages outside of these limits or with unusual values judged to be coding or keying errors were deleted from the quantification analysis for that particular supplement. Subjects who did not know the amount per pill (2,772 cases) were assigned the following default values based on the most common responses for those nutrients from earlier surveys: vitamin A, 1, IU; vitamin C, 25 mg; vitamin E, 1 IU; and calcium, 5 mg. Valid ranges were also established for number of pills per day. Codes outside these limits (affecting a range of three records for vitamin A to 7 records for multiple vitamins), as well as those with unknown values (155 cases), were assigned one pill per day. The data reported here are descriptive in nature, and patterns of supplement use are the emphasis. Multivariate analyses were not performed. Analyses were performed using Statistical Analysis Systems (21). Estimates of the proportion who consumed supplements and average daily nutrients from supplements were calculated using sample weights provided by the National Center for Health Statistics. When appropriate, comparisons of the proportions who used supplements were performed using the normal deviate, Z, and take into account the complex sample survey design in developing variance estimates (22). RESULTS Table 1 shows the distribution of the sample and percent consuming supplements every month daily by age, race, and sex. With few exceptions, women were more likely to consume supplements regularly than were men of the same age and race. Daily usage among white women peaked within the age ranges years (39.9 percent) and years (38.4 percent) and declined slightly at age 75 years or more (34.9 percent). Similarly, within other sex-race categories, usage tended to increase with age, although this pattern was less consistent for black and Hispanic men. Among age-sex categories, whites tended to consume the most supplements compared with all other races. Hispanics appeared to consume supplements at a rate intermediate between those of whites and blacks. Table 2 presents the percent of subjects in user definition categories by sex and race. Of the total population, 51.1 percent stated that they had consumed any supplement in the past year. As the definition for user became more restrictive, that is, as more regular users were sought, the percent of subjects using supplements decreased, leaving only 23.1 percent consuming at least one supplement on a daily year-round basis. Table 3 shows the percentage of subjects with daily usage of any or specific supplements by various demographic and personal characteristics. Of individual supplement types, multivitamins were the most commonly consumed (17.4 percent). In the total population, multivitamin usage was followed, in decreasing order of daily usage, by vitamin C (7.6 percent), calcium (6.2 percent), vitamin E (4.1 percent), and vitamin A ( percent). The statistical significance shown is that of the difference between the proportion in question and that of a selected reference group which, in most cases, represents the lowest proportion of the demographic category (with exceptions noted). Demographic data show that, except for vitamins A and E, high school graduates and college-educated individuals are significantly more likely than those who did not finish high school to consume supplements. Income data show a clear tendency for use of all supplements to increase with income, with subjects having the lowest annual family income (less than $5,) significantly less likely than all other income groups to

4 194 SUBAR AND BLOCK TABLE 1 Survey respondents and percent reporting daily supplement use by age, race, and sex, 1987 National Health Interview Survey Age (years) Total 12.3* ()t 888* 15. (.9) 1, (1.1) 1, (1.4) () (1.7) (2.1) (.6) 7,261 Whites 14.8 (1.4) 1, (1.1) 2, (1.1) 1, (1.6) 1, (1.4) 1, (1.4) 1, (1.7) 1, (.6) 9, (2.6) (2.2) (2.3) (4.1) (4.3) (3.9) (4.9) (1.3) 1,8 Blacks 11.9 (2.2) (1.9) (2.2) (3.7) (3 8) (2.8) (5.) (1.1) 1, (3.2) (2.1) (3 ) (3.7) (4 9) 57 2 (8.3) (8.1) (14) 594 Hispanics 17 7 (4.3) (2.1) (3.) (6.) (4.6) (9.3) (9.5) (1.9) (3.1) (3.6) (3.8) (6.8) (1) (13.3) 11. (.) (1.9) 225 * Weighted percent reporting daily supplement use. t Standard error of the proportion, taking into account sample weights and sample design. X Number of survey respondents. TABLE 2 Other races 8. (3 8) (6.9) (5.8) (5.9) (7.9) (9.9) 14. (.) (2.6) 269 Percent of subjects, by race and sex, consuming any supplement, by frequency of intake categories, 1987 National Health Interview Survey* Frequency of intake Anytime in past year Past month 2 days Past month 2 days Every month Every month 13 days Every month daily Total population Whites Blacks Hispanics Other races * Frequency of use of "any supplement" was assigned the rate of the most frequently consumed supplement type be supplement consumers. Similarly, significantly fewer individuals with incomes below the poverty line were users compared with those above. The demographic factor of region of residence was highly associated with supplement use. For nearly all supplement types except vitamin A, those living in the West were significantly more likely

5 VITAMIN SUPPLEMENT USE IN THE UNITED STATES 195 to use supplements, followed by those in the Northeast. Individuals living in the South were usually the lowest users, but they differed little from those living in the Midwest. Differences by alcohol consumption were less dramatic than for other descriptive variables. Although subjects in the heaviest alcohol consumption category (more than 14 drinks per week) were the lowest supplement users, usage among those in other categories of drinking appears relatively similar for all supplement types. By smoking status, former smokers tended to be the heaviest supplement users, followed closely by never smokers, and current smokers who reported significantly lower use. Individuals in the highest body mass index quartile were the lowest users of any supplement and of multivitamins. Supplement use by response to the question, "Do you believe that diet affects disease?" showed a highly significant trend for those who answered yes to be users of all supplement types compared with those who answered no. Consumers of individual supplement types had age and sex characteristics similar to those seen for intake of "any supplement," with the exception of calcium (data not shown). For age categories above 24 years, women consumed daily calcium supplements at a rate 2-5 times that of men. While among older men there was a moderate increase in the proportion who consumed calcium on a daily basis, this increase among women jumped sharply from 8.6 percent in the age group years to 13.7 percent in the age group years and then to 18.6 percent in the age group years. Table 4 presents data by occupational status regarding daily usage of any supplement. Twenty-two percent or more of individuals in traditional "white-collar" professions (executives, professionals, sales, etc.) were regular supplement users, with the highest rates in senior categories such as executives and those with a professional specialty. Fewer than 2 percent of "bluecollar" workers tended to consume supplements daily. working in the home are included in the category "not in the labor force," and 27.5 percent consume supplements regularly. Earlier tables have addressed the proportion of the population consuming supplements. Table 5 presents quantitative data regarding average daily amounts of nutrients obtained from supplements, both for the total population and for users of supplements only, based upon estimated frequency of consumption during the past year. These data include individuals who consumed supplements as little as once in the past year as well as those who consumed supplements daily. As a point of reference, values for the US Recommended Daily Allowances are given. The US Recommended Daily Allowances are used as standards for purposes of labeling of food and supplements. The US Recommended Daily Allowance found in table 5 is the one that is most commonly used and refers to adults and children above age 4 years (23). Multivitamins of the one-a-day variety typically contain all nutrients at 1 percent of the US Recommended Daily Allowance. Because they are the most commonly consumed supplements, the nutrient content of multivitamins strongly influences the values presented in table 5. For the total population, daily intake of nutrients from supplements did not exceed a value of zero until the 75th percentile (9th percentile for calcium), reflecting the fact that most people do not use supplements. At the 75th percentile, daily intakes for nutrients were below the US Recommended Daily Allowances for all nutrients except vitamin C. At the 9th percentile, however, most nutrient intake was at the US Recommended Daily Allowance level, except for vitamin C and calcium, indicating that 1 percent of the population met or exceeded this standard for those nutrients from supplements. At the 95th percentile, intakes exceeded the US Recommended Daily Allowances for vitamin A, thiamin/riboflavin, vitamin C, and vitamin E, reflecting use of stress-type multivitamins

6 196 SUBAR AND BLOCK TABLE 3 Percent reporting daily supplement use by demographics, 1987 National Health Interview Survey Total Sex f Race White Blackt Hispanic Other Age (years) 17-24f :75 Education (yeare) -8t Any college Annual income (dollars) <5,t 5,-9,999 1,-24,999 25,-49,999 2:5, Poverty index Above poverty Below poverty! Region Northeast Midwest Southf West Alcohol use (drinks/week) Nondnnker <1 2:1-7 >7-14 >14J Cigarette smoking status Never Former Currentt No examined 22,8 9,16 12,92 17,74 3,58 1, ,84 5,332 4,358 2,626 2,684 2,498 1,742 2,458 2,847 8,292 8, ,759 2,428 6,76 6,48 2,247 2,478 17,572 2,914 1,594 4,647 5,626 7,371 4,436 6,823 5,258 5,33 1,88 1, ,48 4,641 6, Population estimate (millions) Any supplement *** 24.8"* *** 22.1"* 26.1*" 32.2'** 31.7"* 31.*** *** 26 5"* *** 22.3"* 24.1*" 26.8*" **' *** 21.4* *" 23 4* 23.9** 23 7* "* 26 "* Multiple type *** 18.5*** '" 17."* 18.7*" 22 4*" 21 6"* 21 9'" *** 2 9*" *** 16.4*** 18.3"' 2.9*" *" *" *** * * 19.7*** Percent who are daily users A "* "* 17"* 2 "* 2."* 17*** Lit ** L4t * 1.1 C *** 8.3*** '" 1.1*" 1.8"* *" 9 7*" " 9.3*** * 7.4*" 7.9*" 1.*** "* "* "* *** E " 4 5"* *** 6.3"* 7.4*" 7."* 6.9*" 3.7t ** 4.1*" 4.*" 5.5*** "* * '" *** Calcium *" 7.1*" " 5.* "* 8 3"* 12.1"* 11.9"* 8 9*" "* 7.2*** * 6 1*" 6.3*** 7.7*** *** * *** 6.5*** 7.*" 6.2*** 5.7* "* 7 '" 41 46

7 VITAMIN SUPPLEMENT USE IN THE UNITED STATES 197 TABLE 3 continued Body mass index quartilei t Believe diet affects disease Yes No/don't know No examined 5,36 5,312 5,442 5, ,33 3,777 Population estimate (millions) Any supplement 24.1*** 25.2*** 23.7** '" 17.4 Multiple type 18 8*** 19 5*" 17.3** *" 13.1 Percent who are daily users A 1.6* " 8 C '*' 4.6 E 3.9$ *" 2.6 Calcium 6. 7."* 6.6* *** 4.4 *p <.5, **p <.1; ***p < 1 p values for the comparison of a given percent with the lowest percentile group (reference group) excluding "unknowns"; exceptions only when noted t Reference group. t Reference group does not represent the lowest percent. Body mass index The cutpoints for men are 22.9, > , > , and >27.3 kg/m 2. The cutpointe for women are S2.9, > , > , and >26.9. TABLE 4 Use of any supplement daily by occupation, 1987 National Health Interuiew Survey^ Occupation White-collar Executive/administrative/managerial Professional specialty Technicians and related support Sales Administrative, including clerical Private household Blue-collar Protective service including police/firefighters Service, not protective or household Farming/forestry/fishing Precision production/craft and repair Machine operators/fabncators/laborers Transportation/material moving Handlers/equipment cleaners/helpers/laboreret Other Military Not in labor force 1,739 2, ,56 2, , ,412 1, , % reporting daily use 25 2*** 26.6*** 23 2*** 24 2**' 22.7"* 27 5** 21* 19.7*" *" 23.7 * p <.5; ** p <.1; *" p <.1 p values for the comparison of a given percent with lowest percentile group (handlers, equipment cleaners, etc.) excluding military (because of small sample size) and unknown. t Standard occupational classification. t Reference category. Includes homemakers. and single-nutrient supplements in a small segment of the population. It should be noted, however, that intake of calcium reached only half of the US Recommended Daily Allowance at the 95th percentile. That is, only 5 percent consumed this much or more. Table 5 also shows the distribution of

8 198 SUBAR AND BLOCK TABLE 5 Distribution of average daily intake of nutrients from supplements among total population and among supplement users, 1987 National Health Interview Survey A (IU) Thiamin/ riboflavin (mg) C (mg) D (IU) E (IU) Iron (mg) Calcium (mg) USRDA* 5, , Total population (n» 22,8) 6th 75th 2, th 5, th 5, nt 7, ,269 9,823 7,571 8,981 8,269 2,764 Percentiles 1th th 1, Supplement users 5th 5, th 5, th 1, , 95th 1, , ,6 * US RDA, Umted States Recommended Daily Allowances for children more than age 4 years and adults (22); based on 1968 Recommended Dietary Allowances (23). t For thiamin, nboflavin, vitamin D, and iron, estimates of number of users and amount of use are based only on multivitamin/mineral use; respondents were not asked about use of these as single supplements For calcium, estimate is based only on use of single supplements. For vitamins A, C, and E, estimates are based on combined use of single supplements and multivitamin/mineral use intake from supplements when the population is restricted just to those who use supplements. Below the 5th percentile (i.e., for almost half of those who report ever using supplements), average daily intake from supplements remained below the US Recommended Daily Allowance for all nutrients. Only at the 9th and 95th percentiles among supplement users did average daily nutrient intakes from supplements other than vitamin C reach levels substantially exceeding the US Recommended Daily Allowance. At the 95th percentile of supplement users, vitamin C and vitamin E intakes were approximately 17 and 14 times the US Recommended Daily Allowance, respectively, while thiamin and riboflavin intakes were approximately seven to eight times this recommendation. A, vitamin D, iron, and calcium intakes were approximately double the US Recommended Daily Allowance or less at the 95th percentile, indicating that few individuals obtain from supplements intakes of these nutrients which are greatly beyond that which could easily be consumed in foods. DISCUSSION The data presented are analyses of current vitamin and mineral supplement practices from a representative national survey. The findings that whites, women, older individuals, and those living in the West are more likely than their counterparts to be supplement consumers are common to other studies of both representative (16-19) and nonrepresentative (7, 8, 1, 13, 15) samples, suggesting that the demographic pattern of usage has not changed substantially in the past 2 decades. Few prior national data have been available regarding supplement use in individuals beyond age 74 years. Like the 1986 NHIS (19), we show a slight but nonsignificant decline in supplementation for this elderly group compared with those in the previous age decade. The 1987 NHIS is also one of the first national studies to address specifically supplement use in the Hispanic population. Our results suggest that Hispanics' pattern of supplementation is intermediate between the rates of whites and blacks. Our results showing that 23.1 percent of subjects take a supplement daily are comparable with those found in National Health and Nutrition Examination Survey I (NHANES I) (21.4 percent) (17) and National Health and Nutrition Examination Survey II (NHANES II) (22.8 percent) (16) for an identical definition of frequency of usage. A study conducted by the Food and

9 VITAMIN SUPPLEMENT USE IN THE UNITED STATES 199 Drug Administration in 1983 (18) and national data from the 1986 NHIS (19) conducted in collaboration with the Food and Drug Administration showed a 39.9 and 36.4 percent rate of usage, respectively. However, it should be noted that in those reports, usage was defined as consumption at least once in the prior 14 days. This definition is approximated in our data by "past month >2 days" for which we found a similar rate of 38.7 percent. It appears, therefore, that both the pattern of supplement use and the frequency of intake have not changed appreciably since the early 197s. Further, our results show the impact of different criteria for defining supplement users on the resulting estimates of rates of usage. Clearly, estimates in which supplement use is defined as "anytime in the past year" grossly overstate the proportion of the population consuming supplements on a daily basis, the group most likely to experience health effects. For example, data from two studies conducted in seven western states (12,15) found usage rates as high as 67.7 and 54 percent, respectively. The definition of usage in the first study (12), however, consisted of any use in the past 2 years, while the definition in the second study (15) consisted of use anytime in the past week. We show in our data that using the definition, "past month >2 days", compared with the more stringent definition from NHANES I and NHANES II (daily use) results in a difference of nearly 16 percent in the number of individuals called "regular users." Compared with national data from NHANES I (17) for individual supplements, our results show no or slight increases in daily use for all supplement subtypes (except calcium); all are within two percentage points of the NHANES I levels: multivitamins (17.4 vs percent), vitamin A ( vs. percent), vitamin E (4.1 vs. 3.7 percent), and vitamin C (7.6 vs. 6. percent). Daily use of calcium, however, increased from percent in NHANES I to 6.2 percent in our study for the total population, and from 1.6 to 9.5 percent for women. This increase, compared with other nutrients, most likely reflects the recent concern regarding osteoporosis in postmenopausal women. Demographic data regarding education, income, and occupation indicate strong socioeconomic influences upon whether individuals take supplements. We show that supplement use increases with both years of education and income for all supplement types. This result, found in past research (12, 16-19), indicates that the pattern of usage has not changed recently. Data from past research has shown that occupation is associated with health behaviors such as increased smoking rates among blue-collar workers (24), and this is consistent with lower supplementation use among bluecollar workers in our research. Another area which appears to influence supplementation is health habits and beliefs and the sociodemographic variables which influence them: 1) Individuals who believe diet affects disease are significantly more likely to be regular supplement users than those who do not; 2) Nondrinkers and lighter drinkers are somewhat more likely to use supplements than are heavier drinkers; 3) Never and former smokers are more likely to be supplement users than are current smokers; and 4) Individuals in the highest quartile of body mass index are less likely to use supplements than are those in the other three quartiles. Data from NHANES I (17) found similar results by alcohol intake, smoking, and body mass index. Our estimates of nutrient intake from supplements are lower at each percentile than are data from the 1986 NHIS (19) which collected detailed label information on supplements consumed in the past 2 weeks. When comparable definitions are used, however, the results obtained from the two surveys are similar. Our data reflect all usage, be it once a year or once a month. When we limited the sample to those who consumed a supplement 2 or more days in the last month, in order to obtain a usage definition comparable with the 1986 NHIS definition of at least once in the past 2

10 11 SUBAR AND BLOCK weeks, less frequent use was eliminated and our values are comparable with the 1986 NHIS data (except for calcium, which is higher in the 1987 NHIS) (data not shown). While collecting data from labels is, no doubt, more accurate for quantification of use over a short period of time, our data are reasonably comparable, are easier to collect and code, and provide useful data regarding "usual" supplement intake over the past year. We found that, with the exception of vitamin C, most Americans, even those at the 9th or 95th percentiles of daily intake from supplements, receive nutrient quantities which do not greatly exceed their Recommended Dietary Allowances (25). C, however, is consumed in quantities greatly above the Recommended Dietary Allowance for adults, but there is little evidence indicating that such doses are harmful (26). Even among supplement users, intake up to the 75th percentile (again, except for vitamin C) does not exceed the Recommended Dietary Allowance or US Recommended Daily Allowance levels. Only at the 9th and 95th percentiles do individuals begin to obtain any higher doses. A is known to be toxic if consumed in intakes of more than 5, IU over a period of months or years for adults (27). Our data show that even among supplement users, intakes at the 95th percentile level (1,814 IU) are approximately one fifth of this and therefore are not likely to reach toxic levels. Intakes of thiamin and riboflavin reach approximately eight times the Recommended Dietary Allowance; however, these nutrients are not known to be toxic at these levels. E, too, although reaching 43 IU daily at the 95th percentile among supplement users, is not known to be toxic even at much higher levels (27). Intakes of iron and calcium at the 95th percentile are not excessive and do not appear to pose any significant health risks. A recent report from the National Academy of Sciences (28) recommends that individuals avoid taking supplements in excess of the Recommended Dietary Allowance. Our data show that few Americans exceed this level, except with vitamin C, and that the risk appears minimal. This does not imply, however, that there are no individuals consuming amounts of nutrients, especially vitamin A, which are known to be toxic. These individuals, however, do not represent great numbers as there were only 18 individuals in this survey (.8 percent of the adult population) with a calculated daily vitamin A intake equal to or greater than 5, IU. Even among these 18 individuals, a close examination suggested that for several of them the estimates might be erroneously high. This was because their high estimated intake was based on a reported daily ingestion of 4-6 "therapeutic"-type multivitamin pills, sometimes in addition to daily use of high dosage single vitamin A supplements. We feel that such multivitamin use might reflect products currently on the market in which several tablets must be consumed in order to provide nutrient dosage equal to a single high-dosage multivitamin pill, and thus represented overestimates of actual dose. Although we are suspicious that the data for approximately half of these 18 individuals may be erroneous, we chose not to exclude them. Health professionals should be concerned about individuals ingesting toxic amounts of nutrients; however, it is clear that they represent a very small segment of the adult population. We have reported here that the proportion of the population consuming supplements on a daily basis has changed very little since the early 197s, and we have illustrated that estimates of the proportion who are regular supplement users are very sensitive to the definition of "regular," as well as to the demographic characteristics of the population sampled. When intake over an entire year is considered, relatively few Americans consume supplements at nutrient levels substantially in excess of the Recommended Dietary Allowance. Finally, we have shown that a frequency-type method of data collection can provide quan-

11 VITAMIN SUPPLEMENT USE IN THE UNITED STATES 111 titative estimates of nutrient intake from supplements that are comparable with those obtained by more detailed methods. Currently, scientists are seeking to determine whether nutrients are protective for chronic diseases such as cancer and heart disease. Supplementation has a substantial impact on the intake of vitamins and minerals for some individuals and must be assessed in research seeking to determine the dietary factors associated with disease. Use of vitamin and mineral supplements is a health habit that is not likely to disappear, and a need exists to continue to assess its health impact and to regard it as an important source of total nutrient intake. REFERENCES 1 Willett WC, MacMahon B. Diet and cancer an overview (Part 1). N Engl J Med 1984; Willett WC, MacMahon B. Diet and cancer an overview (Part 2) N Engl J Med 1984, Block G. C and cancer prevention, the epidemiologic evidence Am J Clin Nutr (in press) 4. Garry PJ, Goodwin JS, Hunt WC, et al. Nutritional status in a healthy elderly population' dietary and supplemental intakes Am J Clin Nutr 1982; Garry PJ, Goodwin JS, Hunt WC, et al. Nutritional status in a healthy elderly population: vitamin C Am J Chn Nutr 1982; Gray GE, Paganini-Hill A, Ross RK. Dietary intake and nutrient supplement use in a southern California retirement community. Am J Clin Nutr 1983;38: Read MH, Graney AS Food supplement usage by the elderly. J Am Diet Assoc 1982;8: Schneider CL, Norlund DJ Prevalence of vitamin and mineral supplement use in the elderly. J Fam Pract 1983;17: Bootman JL, Wertheimer Al. Patterns of vitamin usage in a sample of university students. J Am Diet Assoc 198;77: Bowerman SJA, Hamll I. Nutrient consumption of individuals taking or not taking nutrient supplements. J Am Diet Assoc 1983;83: Hamll I, Bowski MM. Relationship of age and sex to nutrient supplement usage in a group of adults in Colorado. J Nutr Elderly 1981;l: Schutz HG, Read M, Bendel R, et al Food supplement usage in seven western states. Am J Chn Nutr 1982; Willett W, Sampson L, Bain C, et al supplement use among registered nurses. Am J Clin Nutr 1981;34: Worsley A, Crawford DA, Baghurst KI. Diet, food beliefs and dietary supplementation. Ecology Food Nutr 1985;19: Medeiros DM, Bock MA, Ortiz M, et al. and mineral supplementation practices of adults in seven western states J Am Diet Assoc 989; Koplan JP, Annest JL, Layde PM, et al. Nutrient intake and supplementation in the United States (NHANES II) Am J Public Health 1986,76: Block G, Cox C, Madans J, et al. supplement use, by demographic characteristics Am J Epidemiol 1988;127: Stewart ML, McDonald JT, Levy AS, et al /mineral supplement use- a telephone survey of adults in the United States. J Am Diet Assoc 1985, Moss A, Levy AR, Kim I. Use of vitamin and mineral supplements in the US: current users, types of products, and nutrients Hyatteville, MD- National Center for Health Statistics, Advanced data, vital and health statistics (no. 174) (DHHS publication no ) 2. National Center for Health Statistics, Public Health Service, US Department of Health and Human Services. The National Health Interview Survey design and procedures Hyattsville, MD Department of Health and Human Services, (Series 1, no. 18) (DHHS publication no. (PHS) ) 21. SAS Institute, Inc SAS user's guide. Cary, NC: SAS Institute, Inc, Shah BV. SESUDAAN: Standard errors program for computing of standardized rates from sample survey data. Research Triangle Park, NC: Research Triangle Institute (unpublished report) 23. Food and Drug Administration, Department of Health and Human Services. Nutrition labeling of food Code of Federal Regulations, Title 21, Part Washington, DC. US GPO, Centers for Disease Control, Office on Smoking and Health, Public Health Services, US Department of Health and Human Services. Reducing the health consequences of smoking, 25 years of progress. A report of the Surgeon General, Rockville, MD: Department of Health and Human Services, (DHHS publication no. (CDC) ). 25 Committee on Dietary Allowances Recommended dietary allowances. Washington, DC. National Academy of Sciences, Block G, kes M. Ascorbic acid in cancer prevention In. Moon TE, Micozzi MS, eds. Diet and cancer prevention. Investigating the role of micronutrients New York: Marcel Dekker, Inc, 1989: Committee on Safety Toxicity and Misuse of s and Trace Minerals, National Nutrition Consortium, Inc. -mineral safety, toxicity and misuse Chicago: American Dietetic Association, National Research Council. Committee on Diet and Health, Food and Nutrition Board. Commission on Life Sciences. Diet and health. Implications for reducing chronic disease risk. Washington, DC: National Academy Press, 1989.