Method Used to Set the Adequate Intake

An Adequate Intake (AI) is used as the goal for intake by infants.

Ages 0 through 6 Months. The AI reflects the observed mean thiamin intake of infants consuming human milk. Thus, the thiamin AI for young infants is based on mean intake data from infants fed human milk exclusively during their first 6 months and uses the thiamin concentration of milk produced by well-nourished mothers. There are no reports of full-term infants who were exclusively fed milk from U.S. or Canadian mothers who manifested any signs of thiamin deficiency; however, infants breastfed by mothers with beriberi have been reported to develop beriberi themselves by age 3 to 4 weeks (Hytten and Thomason, 1961). The thiamin content of human milk was similar for well-nourished mothers who received vitamin supplements and for those who did not (Nail et al., 1980; Pratt and Hamil, 1951).

The thiamin concentration is low in colostrum (approximately 0.01 μg/L). The mean concentration of thiamin in mature human milk is 0.21 ± 0.04 mg/L (mean ± standard deviation) (Committee on Nutrition, 1985). Using the mean volume for intake of human milk of 0.78 L/day (see Chapter 2) and the average thiamin content of 0.21 mg/L, the AI for thiamin is 0.16 mg/day for infants ages 0 through 6 months, which is rounded to 0.2 mg. For the reference infant weight of 7 kg, this corresponds to 0.03 mg/kg/day.

Blood concentration of total thiamin (phosphorylated and nonphosphorylated) has been shown to decrease with age: in a cross-sectional study of well-nourished individuals, blood thiamin concentrations in infants less than 3 months of age (n = 64) averaged 258 ± 63 nmol/L (75 ± 23 μg/L) (mean ± standard deviation), infants 3 to 12 months of age (n = 100) averaged 214 ± 44 nmol/L (64 ± 13 μg/L), while in children and young adults (n = 159) the value decreased to 187 ± 39 μmol/L (56 ± 12 μg/L) (Wyatt et al., 1991). Because total thiamin concentrations in whole blood and cerebrospinal fluid decrease in the first 12 to 18 months of life, age-specific norms should be used for determining thiamin status in infancy.

Ages 7 through 12 Months. If the reference body weight ratio method described in Chapter 2 to extrapolate from the AI for thiamin for infants ages 0 through 6 months is used, the AI for thiamin for the older infants would be 0.2 mg/day after rounding. The second method (see Chapter 2), extrapolating from the Estimated Average Requirement (EAR) for adults and adjusting for the expected variance to estimate a recommended intake, gives an AI of 0.3 mg of thiamin, a value higher than that obtained from the first method.

Alternatively, the AI for thiamin for infants ages 7 through 12 months could be calculated by using the estimated thiamin content of 0.6 L of human milk, the average volume consumed by this age group (thiamin content equals 0.13 mg), and adding the amount of thiamin provided by solid foods (0.5 mg), as estimated by Montalto et al. (1985) (see Chapter 2). The result equals approximately 0.6 mg/day. This value was judged to be unreasonably high because it is two to three times the extrapolated values given above. Thus the AI for thiamin is 0.3 mg/day for infants ages 7 through 12 months—the value extrapolated from estimates of adult requirements.

Thiamin AI Summary, Ages 0 through 12 Months

Method Used to Estimate the Average Requirement

No direct data were found on which to base an EAR for children ages 1 through 8 years. In the absence of additional information, EARs and Recommended Dietary Allowances (RDAs) for these age groups have been extrapolated from adult values by using the method described in Chapter 2.

Thiamin EAR and RDA Summary, Ages 1 through 8 Years

The RDA for thiamin is set by assuming a coefficient of variation (CV) of 10 percent (see Chapter 1) because information is not available on the standard deviation of the requirement for thiamin; the RDA is defined as equal to the EAR plus twice the CV to cover the needs of 97 to 98 percent of the individuals in the group (therefore, for thiamin the RDA is 120 percent of the EAR).

Evidence Considered in Estimating the Average Requirement

Five studies were found for this age group, none of which involved children younger than 13 years. In an observational study of 19 boys and 35 girls aged 13 or 14 years, thiamin intake was calculated from 7-day food records and was also analyzed by high-performance liquid chromatography from duplicate portions (Bailey et al., 1994). The correlations of the results from the food records and the analyses of duplicate portions were significant but moderate (r = 0.59 for boys and 0.43 for girls). However, the indicators of thiamin status (erythrocyte transketolase, erythrocyte transketolase activity coefficient, and total erythrocyte thiamin concentration) were not correlated with each other. Moreover, none of them was correlated with thiamin intake as estimated from the food records or measured in the duplicate portions. A substantial percentage of the subjects (girls, 12 percent; boys, 17 percent) had activity coefficients that indicated a high risk of thiamin deficiency according to Brin's criterion for adults (Brin, 1970) even though estimated intakes were above 0.4 mg/1,000 kcal.

A controlled-diet, dose-response experiment was conducted with nine girls aged 16 to 18 years to examine the thiamin requirement (Hart and Reynolds, 1957). In this study, the girls were given 0.29 mg of thiamin/1,000 kcal/day (0.63 mg/day total) for the first 16-day period and 0.6 mg/1,000 kcal/day (1.3 mg/day) for the second 16-day period. The adequacy of intake was assessed by measuring total daily thiamin excretion, the percentage of consumed thiamin that was excreted, the ratio of thiamin to creatinine in the urine, and the percentage of excretion of a 5-mg oral test dose of thiamin hydrochloride. Using a modification of the thiochrome method for thiamin determination, the investigators were unable to obtain reliable measurements of the amount of thiamin excreted on the low-thiamin diet. The authors noted that the subjects became irritable and uncooperative and lost the ability to concentrate when fed the low-thiamin diet—symptoms also noted by others in the early stage of thiamin deficiency. On the diet that provided 1.3 mg/day of thiamin, 24-hour thiamin excretion ranged between 0.27 and 0.44 μmol (81 and 133 μg). These data suggest that the average thiamin requirement is less than 1.3 mg/day, especially considering the short period of repletion and the use of a generous cutoff point for urinary thiamin excretion, but they do not allow further refinement of the estimate.

In a study of eight boys aged 14 to 17 years, Dick and colleagues (1958) calculated thiamin requirements from a regression of excretion on intake at five levels of thiamin that ranged from 0.6 to 2.7 mg/day. By taking the abscissa of the intersection of two straight lines fitted to the observations on each subject, a mean requirement of 1.41 mg/day was computed. However, at this level of intake, mean urinary excretion averaged 0.618 μmol/day (186 μg/day)—a value far in excess of usual cutoffs.

In the absence of additional definitive information about requirements, EARs and RDAs for thiamin for these age groups were extrapolated from the adult values by using the method described in Chapter 2. Because only urinary excretion of thiamin was measured, the results reported by Hart and Reynolds (1957) are not considered strong enough to warrant adjustment of results from the extrapolation method.

Thiamin EAR and RDA Summary, Ages 9 through 18 Years

The RDA for thiamin is set by assuming a coefficient of variation (CV) of 10 percent (see Chapter 1) because information is not available on the standard deviation of the requirement for thiamin; the RDA is defined as equal to the EAR plus twice the CV to cover the needs of 97 to 98 percent of the individuals in the group (therefore, for thiamin the RDA is 120 percent of the EAR).

Indicators Used to Estimate the Average Requirement

It was necessary to review data from studies that used various indicators of thiamin sufficiency to derive an EAR for thiamin for adults. In reviewing the studies (Table 4-2), heavy weight was given to the carefully controlled, thiamin depletion-repletion experiment conducted with seven healthy young men (age not specified) in a metabolic unit (Sauberlich et al., 1979). The investigators concluded that thiamin at 0.30 mg/1,000 kcal (approximately 1.0 mg/day) met the minimum requirement for young men as determined by using urinary excretion of thiamin, and it appears that this value is close to the average requirement for normal erythrocyte transketolase activity. This value is slightly lower than the 1.2 and 1.0 mg total values designated by Anderson and colleagues (1986) as minimal for men and women, respectively, determined by using erythrocyte transketolase activity.

TABLE 4-2

Metabolic Studies Providing Evidence Used to Derive the Estimated Average Requirement (EAR) for Thiamin for Adults.

Studies by Bamji (1970) and Ziporin and coworkers (1965) support the conclusion that 0.30 mg/1,000 kcal/day is a minimum thiamin requirement to prevent overt signs and symptoms of deficiency, although at this level urinary excretion of thiamin was abnormal in some subjects. Studies by Kraut and colleagues (1966) and Reuter et al. (1967) suggest a much higher requirement. Although intakes of approximately 0.7 mg/day (0.4 mg/1,000 kcal) were found to meet the minimum requirement for thiamin based on erythrocyte transketolase activity (Reuter et al., 1967), the achievement of maximum erythrocyte transketolase activity required intakes of 2.0 to 2.5 mg/day (Kraut et al., 1966; Reuter et al., 1967). Several studies indicated that intakes of 0.075 to 0.29 mg/1,000 kcal lead to severe irritability and other symptoms and signs of deficiency (Foltz et al., 1944; Horwitt et al., 1948; Wood et al., 1980). Data are not sufficient to indicate differing requirements for adults 19 through 30 versus 31 through 50 years of age.

Thiamin EAR and RDA Summary, Ages 19 through 50 Years

Examination of the data in Table 4-2 indicates that the EAR for thiamin is at least 0.3 mg/1,000 kcal or 0.8 mg/day and that intakes greater than 1.0 mg are marginally adequate for normal erythrocyte transketolase activity and generally adequate for urinary thiamin excretion. Because of the uncertainties of the dietary intake estimates in the studies by Anderson et al. (1986) and Henshaw et al. (1970), greater weight was given to the well-controlled studies of Sauberlich et al. (1979). With the assumption of a curvilinear relationship with increasing intake, it is concluded that the EAR for thiamin is 1.0 mg/day for men and 0.9 mg/day for women, which represents about a 10 percent decrease for women based on body size and energy needs.

The RDA for thiamin is set by assuming a coefficient of variation (CV) of 10 percent (see Chapter 1) because information is not available on the standard deviation of the requirement for thiamin; the RDA is defined as equal to the EAR plus twice the CV to cover the needs of 97 to 98 percent of the individuals in the group (therefore, for thiamin the RDA is 120 percent of the EAR).

Evidence Considered in Estimating the Average Requirement

Several studies of the thiamin status of the elderly have been conducted but provide little direct information on which to base nutrient requirements. Laboratory indicators of status suggest that a substantial percentage (20 to 30 percent) of the population has values suggestive of deficiency, but reported intake is not correlated with the laboratory results. Nichols and Basu (1994) investigated the relationship between thiamin intake and the thiamin pyrophosphate (TPP) effect in a group of 60 randomly selected, free-living elderly men and women in Alberta, Canada, aged 65 to 74 years. Thiamin intake estimated from three nonconsecutive food records was not significantly correlated with the TPP effect. In this study, only 57 percent of the subjects were described as having adequate thiamin status (TPP effect of less than 14 percent) even though the reported mean thiamin intake was 1.7 ± 0.12 mg (standard error)/day for men and 1.4 ± 0.01 mg/day for women.

O'Rourke and coworkers (1990) reported a lower TPP effect for 10 healthy elderly individuals (aged 70 to 82 years) than for 13 institutionalized elderly (aged 67 to 92 years) who were apparently free of significant gastrointestinal, hepatic, or renal disease. No differences were found in erythrocyte transketolase activity or erythrocyte thiamin content. The investigators did not report on relationships between thiamin intake and the indicators of status.

In a study of 75 elderly Finnish women and men (aged 50 to 94 years), some of whom were institutionalized, slightly over 20 percent were described as having marginal thiamin deficiency defined as greater than 15 percent TPP (Pekkarinen et al., 1974). The mean intake of thiamin was much lower than in the Nichols and Basu (1994) study mentioned above and it varied by group, ranging from 0.46 to 0.79 mg/day in women and from 0.60 to 0.84 mg/day in men. The enzyme activity of the erythrocytes was not significantly correlated with thiamin intake as determined by thiochrome analysis of collected food samples.

Similarly, Hoorn and colleagues (1975) reported that 23 percent of 153 geriatric patients aged 65 to 93 years were deficient in thiamin as determined by a transketolase activation coefficient greater than 1.27. Status became normal in all patients after the administration of 20 mg of thiamin daily for 12 days. No dietary information was provided.

A depletion-repletion study of 10 active, healthy elderly women (aged 52 to 72 years, 9 of whom were 63 years or older) measured urinary thiamin excretion after various thiamin intakes (Oldham, 1962). Eight young women aged 18 to 21 years were also studied. On a thiamin intake of less than 0.40 mg/day, the urinary thiamin excretion decreased more quickly for the older than for the younger women. By day 11 or 12, the thiamin excretion of the older women averaged only 0.153 μmol/day (0.05 μg/day), and they complained of fatigue, headaches, and irritability and canceled social engagements. Comparable results were not seen for the younger women until day 19 or 20. When thiamin intake was increased, the older women's urinary thiamin excretion did not increase as quickly as did that of the younger women. The authors concluded that the thiamin requirement of elderly women is higher than that of young women and that the ratio of thiamin to energy must be higher, but the highest thiamin intake level tested, 0.81 mg/day, showed a very wide range of urinary thiamin excretion, especially after 6 days at this intake.

In a randomized, double-blind treatment trial of either 10 mg of thiamin or placebo (Wilkinson et al., 1997), treated subjects who had low TPP concentrations twice when measured before randomization were reported to experience subjective benefits (improved quality of life) and lower blood pressure and weight; those with only one low TPP value prior to randomization did not benefit from treatment.

Although there are some data to suggest that requirements might be somewhat higher in the elderly than in younger adults (e.g., Oldham, 1962), there is also a concomitant decreased energy utilization that may offset this. Further study of this question needs to be conducted. Thus the EAR is assumed to be the same for elderly and younger adults.

Thiamin EAR and RDA Summary, Ages 51 Years and Older

The EAR for thiamin for adults ages 50 and older is set at the same level as for younger adults—1.0 mg/day for men and 0.9 mg/day for women.

The RDA for thiamin is set by assuming a coefficient of variation (CV) of 10 percent (see Chapter 1) because information is not available on the standard deviation of the requirement for thiamin; the RDA is defined as equal to the EAR plus twice the CV to cover the needs of 97 to 98 percent of the individuals in the group (therefore, for thiamin the RDA is 120 percent of the EAR).

Method Used to Estimate the Average Requirement

The few studies of the thiamin need of pregnant women focus mainly on single indicators of status, usually without reference to dietary intake. For example, one measurement of transketolase activity was made in each of 556 pregnant German women at various stages of gestation (Heller et al., 1974). The mean activation coefficient was 1.13 whereas that of a reference group of 300 blood donors was 1.05; the cutoff value of normal activation coefficients, derived from data on nonpregnant adults, was 1.20. Twenty-six percent of the women with uncomplicated pregnancies and 21 percent of those with complications had activation coefficients above the cutoff and were classified as abnormal.

Regardless of the nutritional status of the mother, erythrocyte transketolase activity was higher in cord blood than in maternal blood (Tripathy, 1968). Similarly, the free thiamin concentration was higher in cord blood (Slobody et al., 1949). Transketolase activity in cord blood tended to be proportional to that in maternal blood and higher in the blood of pregnant than of nonpregnant women (Tripathy, 1968). In 103 pregnant Malaysian women whose staple diet was rice, 36 percent had a TPP effect greater than 25 percent—a larger percentage than was found for males and non-pregnant women (Chong and Ho, 1970).

Oldham and coworkers (1950) found a very strong correlation (r = 0.98) between total thiamin intake and excretion but no consistent decrease in thiamin excretion or in the percentage of a test dose excreted over the course of pregnancy. These investigators compared their results with those from studies of nonpregnant women of comparable ages (Daum et al., 1948; Hathaway and Strom, 1946; Oldham et al., 1946) and found that the pregnant women excreted two to three times as much thiamin as did the nonpregnant women on similar intakes (estimated at less than 1 mg) whereas their excretion of a test dose was similar. In contrast, Toverud (1940) observed no or minimal excretion of thiamin in the urine normally or after a load test in 46 percent of 114 pregnant women. Lockhart and coworkers (1943) reported that approximately three times as much thiamin, as obtained from both supplements and diet, was needed by 16 pregnant women to achieve the urinary excretion peak in the tenth lunar month as was needed by a group of nonpregnant women.

Thiamin EAR and RDA Summary, Pregnancy

For pregnancy the requirement is increased by about 30 percent based on increased growth in maternal and fetal compartments (approximately 20 percent) and a small increase in energy utilization (about 10 percent). This results in an additional requirement for pregnancy of 0.27 ≅ 0.3 mg/day of thiamin. Data from the studies cited above are equivocal about the effects of pregnancy on thiamin requirements and thus are not useful in refining this estimate. Adding 0.3 to the EAR of 0.9 mg for nonpregnant, nonlactating women gives an EAR for the second and third trimesters of pregnancy of 1.2 mg. No adjustment is made for the woman's age.

The RDA for thiamin is set by assuming a coefficient of variation (CV) of 10 percent (see Chapter 1) because information is not available on the standard deviation of the requirement for thiamin; the RDA is defined as equal to the EAR plus twice the CV to cover the needs of 97 to 98 percent of the individuals in the group (therefore, for thiamin the RDA is 120 percent of the EAR).

Method Used to Estimate the Average Requirement

For lactating women it is assumed that 0.16 mg of thiamin is transferred in their milk each day when daily milk production is 0.78 L (see “Ages 0 through 6 Months”). To estimate the average thiamin requirement of lactating women, an additional 0.1 mg of thiamin is added to the EAR (0.9 mg/day) for the nonpregnant, nonlactating woman to cover the energy cost of milk production. Thus, the EAR for thiamin for the lactating woman is

Women who are breastfeeding older infants who are eating solid foods might need slightly less thiamin because of a lower volume of milk production.

Thiamin EAR and RDA Summary, Lactation

The RDA for thiamin is set by assuming a coefficient of variation (CV) of 10 percent (see Chapter 1) because information is not available on the standard deviation of the requirement for thiamin; the RDA is defined as equal to the EAR plus twice the CV to cover the needs of 97 to 98 percent of the individuals in the group (therefore, for thiamin the RDA is 120 percent of the EAR).

Adverse Effects

There are no reports available of adverse effects from consumption of excess thiamin by ingestion of food and supplements. Because the data are inadequate for a quantitative risk assessment, no Tolerable Upper Intake Level (UL) can be derived for thiamin. Supplements that contain up to 50 mg/day of thiamin are widely available without prescription, but the possible occurrence of adverse effects resulting from this level or more of intake appears not to have been studied systematically. The limited evidence of adverse effects after large intakes of thiamin is summarized here.

Anaphylaxis. There have been occasional reports of serious and even fatal responses to the parenteral administration of thiamin (Stephen et al., 1992). The clinical characteristics have strongly suggested an anaphylactic reaction. Symptoms associated with thiamin-induced anaphylaxis include anxiety, pruritus, respiratory distress, nausea, abdominal pain, and shock, sometimes progressing to death (Laws, 1941; Leitner, 1943; Reingold and Webb, 1946; Schiff, 1941; Stein and Morgenstern, 1944; Stiles, 1941).

Allergic Sensitivity and Pruritus. Royer-Morrot and colleagues (1992) reported one case of pruritus after an intake of 500 mg/day of thiamin intramuscularly. Another study (Wrenn et al., 1989), which involved intravenous administration of 100 mg of thiamin hydrochloride to 989 patients, reported a burning effect at the injection site in 11 patients and pruritus in 1 patient. No reports of pruritus after thiamin ingestion were found. Because pruritus was only observed with parenteral administration and at a dosage well above the maximum that can be absorbed, it is irrelevant for setting a UL. The finding of Wrenn and coworkers (1989) supports the conclusion that even intravenous administration of high doses of thiamin is relatively safe.

The apparent lack of toxicity of supplemental thiamin may be explained by the rapid decline in absorption that occurs at intakes above 5 mg (Hayes and Hegsted, 1973; SCOGS/LSRO, 1978) and the rapid urinary excretion of thiamin (Davis et al., 1984; McAlpine and Hills, 1941; Najjar and Holt, 1940).