Absorption and Transport

In animal tissue the major form of B₆ is PLP; next is PMP. Absorption in the gut involves phosphatase-mediated hydrolysis followed by transport of the nonphosphorylated form into the mucosal cell. Transport is by a nonsaturable passive diffusion mechanism. Even extremely large doses are well absorbed (Hamm et al., 1979). PN glucoside is absorbed less effectively than are PLP and PMP and, in humans, is deconjugated by a mucosal glucosidase (Nakano and Gregory, 1997). Some PN glucoside is absorbed intact and can be hydrolyzed in various tissues.

Metabolism

Most of the absorbed nonphosphorylated B₆ goes to the liver. PN, PL, and PM are converted to PNP, PLP, and PMP by PL kinase. PNP, which is normally found only at very low concentrations, and PMP are oxidized to PLP by PNP oxidase. PMP is also generated from PLP via aminotransferase reactions. PLP is bound to various proteins in tissues; this protects it from the action of phosphatases. The capacity for protein binding limits the accumulation of PLP by tissues at very high intakes of B₆ (Merrill et al., 1984). When this capacity is exceeded, free PLP is rapidly hydrolyzed and nonphosphorylated forms of B₆ are released by the liver and other tissues into circulation. At pharmacological doses of B₆, the high capacities for PLP-protein binding of muscle, plasma, and erythrocytes (hemoglobin) allow them to accumulate very high levels of PLP when other tissues are saturated (Lumeng et al., 1978).

PLP in the liver can be oxidized to 4-PA, which is released and excreted. The major PLP-binding protein in plasma is albumin. PLP is the major form of the vitamin in plasma and is derived entirely from liver as a PLP-albumin complex (Fonda et al., 1991; Leklem, 1991). Tissues and erythrocytes can transport nonphosphorylated forms of the vitamin from plasma. Some of this is derived from plasma PLP after phosphatase action. In tissues, conversion of the transported vitamin to PLP, coupled with protein binding, allows accumulation and retention of the vitamin. B₆ in tissues is found in various subcellular compartments but primarily in the mitochondria and the cytosol.

Excretion

Normally, the major excretory product is 4-PA, which accounts for about half the B₆ compounds in urine (Shultz and Leklem, 1981). Other forms of the vitamin are also found in urine. With large doses of B₆, the proportion of the other forms of the vitamin increases. At very high doses of PN, much of the dose is excreted unchanged in the urine. B₆ is also excreted in feces but probably to a limited extent (Lui et al., 1985). Microbial synthesis of B₆ in the lower gut makes it difficult to evaluate the extent of this excretion.

Body Stores

Pharmacokinetic analyses of urinary excretion of a tracer dose of labeled PN and its metabolites have suggested a two-compartment model for body B₆ stores (Johansson et al., 1966). With this approach, body stores have been estimated at 365 μmol (61 mg) or 7.7 μmol/kg in a healthy 20-year-old woman and 660 μmol (110 mg) or 8.8 μmol/kg in a 25-year-old man. Overall body half-lives were about 25 days (Shane, 1978). Intake (and excretion) was estimated to be 1.5 mg (9 μmol) for the woman and 3.4 mg (20 μmol) for the man. The two-compartment model has been questioned because muscle stores most of the body's B₆; the pool in muscle appears to turn over very slowly. This fact may have not been considered, resulting in a substantial underestimation of body stores. Coburn and colleagues (1988a) estimated the B₆ content of muscle biopsies and, by assuming that muscle represented 80 percent of the body B₆ store, calculated a total body store of about 1,000 μmol (167 mg).

Extrapolation of data from studies with experimental animals to assess the B₆ requirement for maintenance and growth (Coburn et al., 1987, 1988b) indicates that 1 mg of PN would be an adequate intake for the adult. Modeling of human B₆ pools has also led to an assessment of a minimum requirement of about 0.4 mg/day of PN (Coburn, 1990).

Drug Interactions

Drugs that can react with carbonyl groups have the potential to interact with PLP. Isoniazid, which is used in the treatment of tuberculosis, and L-DOPA, which is metabolized to dopamine, have been reported to reduce plasma PLP concentrations (Bhagavan, 1985; Weir et al., 1991).

Oral Contraceptive Agents

A number of studies have reported decreases in B₆ status indicators in women receiving high-dose oral contraceptives (Rose, 1978; Shane and Contractor, 1975). Plasma PLP concentrations are decreased but the decrease is quite small. Normalization of the tryptophan load test in subjects receiving oral contraceptives requires very high levels of PN, up to 25 mg (Rose, 1978). This probably reflects hormonal stimulation of tryptophan catabolism rather than any deficiency of B₆ per se. These studies were conducted when the level of estrogen in oral contraceptive agents was three to five times higher than it is currently.

Alcohol

Alcoholics have low plasma PLP concentrations, and this reduced B₆ status is distinct from that caused by liver disease or poor diet. Acetaldehyde but not ethanol decreases net PLP formation by cells and is thought to compete with PLP for protein binding. This may make cellular PLP more susceptible to hydrolysis by membrane-bound phosphatase (Lumeng and Li, 1974). The extent to which this causes an increased B₆ requirement is not known.

Method Used to Set the Adequate Intake

An Adequate Intake (AI) is used as the goal for intake by infants. Limited data are available on B₆ requirements of the infant, but an AI can be set based on human milk B₆ content, which varies with maternal B₆ intake. West and Kirksey (1976) reported an average B₆ content of 0.13 mg/L of human milk at maternal B₆ intakes of less than 2.5 mg and 0.24 mg/L at intakes of 2.5 to 5 mg. However, milk B₆ content was quite variable between subjects at similar B₆ intakes. Borschel and colleagues (1986) reported that the milk of mothers receiving 2.5 mg/day of PN as a supplement contained 0.15 to 0.21 mg/L of B₆.

Ages 0 through 6 Months. The AI reflects the observed mean B₆ intake of infants consuming human milk. Thus, the B₆ AI for young infants is based on mean intake data from infants fed human milk as the principal food during their first 6 months and uses the B₆ concentration of milk produced by well-nourished mothers. There are no reports of full-term infants exclusively and freely fed human milk from U.S. or Canadian mothers who manifested any signs of B₆ deficiency. The mean concentration of B₆ in human milk of well-nourished but unsupplemented mothers with intakes near the Recommended Dietary Allowance (RDA) is 0.13 mg/L (West and Kirksey, 1976). Using the mean volume of milk consumed of 0.78 L/day (see Chapter 2) and the average B₆ content of 0.13 mg/L, the AI for B₆ is 0.1 mg/day for infants ages 0 through 6 months.

Little information is available on optimal plasma B₆ concentrations in the infant. Cord vein and artery PLP concentrations are very high in the fetus and in the newborn. They progressively decrease throughout the first year and then more slowly throughout life. Kang-Yoon and coworkers (1995) found lower B₆ content in milk of mothers of preterm infants even when the mothers were receiving 2 mg/day of B₆ as a supplement. Plasma PLP in infants correlated with their B₆ intake. In infants of mothers supplemented with 2 mg/day of PN, plasma PLP fell to 32 nmol/L by 1 month compared with 270 nmol/L in plasma of infants supplemented with 0.4 mg/day of B₆ directly. Formula-fed infants also had much higher plasma PLP values (about 200 nmol/L). In healthy infants, 0.3 mg of B₆ prevented abnormal tryptophan metabolites (Bessey et al., 1957).

In the 1950s convulsions were reported in infants receiving formula in which much of the B₆ content had been destroyed in processing. The convulsions were promptly treated by PN administration (Coursin, 1954); 0.26 mg was sufficient (Bessey et al., 1957) but 1 mg was required to correct abnormal tryptophan metabolites. The B₆ content of the overly processed formula has been estimated to average 0.062 mg/L or less (Borschel, 1995). Deficiency symptoms were not observed in infants receiving other formula preparations containing 0.096 mg/L of B₆, which is equivalent to an intake of 0.075 mg/day (Borschel, 1995). This level of intake is similar to the AI of 0.1 mg suggested from normal intakes and raises questions about whether an AI of 0.1 mg, which appears to be adequate on the basis of normal human milk intakes, may be slightly low. Although it is not possible to set a plasma PLP cutoff value for the infant for adequate status, one would expect it to be at least that used for the adult, if not higher. The drop in plasma PLP concentrations to 32 nmol/L, reported above, in 1-month-old infants of mothers supplemented with B₆ suggests that plasma PLP values are somewhat lower than 32 nmol/L in breastfed infants of mothers who are not supplemented. These data also suggest that the infant fed human milk does not have a large reservoir of B₆.

Other studies, however, have reported higher plasma PLP concentrations in infants of unsupplemented mothers. Andon and colleagues (1989) reported an average human milk B₆ content of 0.12 mg/L in mothers with average dietary intakes of 1.4 mg of B₆. The plasma PLP in the infants of these mothers averaged 54 nmol/L.

A plasma PLP concentration of 15 nmol/L was reported for a full-term infant experiencing B₆-responsive convulsions (Kirksey and Roepke, 1981). This is the only report of a plasma PLP concentration in an infant experiencing convulsions, and the possibility could not be excluded that other factors or a B₆-dependency syndrome contributed to the convulsions.

Ages 7 through 12 Months. If the reference body weight ratio method described in Chapter 2 to extrapolate from the AI for B₆ for infants ages 0 through 6 months is used, the AI 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 for B₆ of 0.4 mg/day, a value that is higher than that obtained from the first method. Because the AI for the young infants appears to be close to the requirement on the basis of studies mentioned in this section, the AI for B₆ is set at 0.3 mg/day for infants ages 7 through 12 months—as the mean obtained from the two types of extrapolation.

B₆ AI Summary, Ages 0 through 12 Months

Special Considerations

Infant formula typically provides much higher levels of B₆ to the infant than does human milk (Borschel et al., 1986), and formula-fed infants have higher PLP concentrations and lower erythrocyte aspartate aminotransferase (α-EAST) values than do infants fed human milk (Heiskanen et al., 1994).

Evidence Considered in Estimating the Average Requirement

Information on the B₆ requirements of children and of adolescents is too limited for an EAR to be established directly. For example, the few studies assessing intakes and 4-pyridoxic acid (4-PA) excretion are not useful for determining a requirement.

A recent study that tracked some B₆ status indicators from children aged 2 months to 11 years (Heiskanen et al., 1995) demonstrated a gradual decrease in erythrocyte PLP with age, similar to that previously reported for plasma PLP (Hamfelt and Tuvemo, 1972), and an increased erythrocyte α-EAST. Interestingly, there was a tendency for those at the lowest percentiles of values to remain so over this period, which could reflect a genetic component, diet, or both. The setting of normal ranges for B₆ status indicators is complicated by age-dependent changes in these indicators.

A few studies have investigated indicators of B₆ status in preschool children and adolescents. Fries and coworkers (1981) measured plasma PLP concentrations in 35 boys and girls aged 3 and 4 years. Average food intakes of B₆ ranged from 0.9 mg in the 4-year-old boys to 1.3 mg in the 3-year-old boys. Some of the children received additional supplements. The average plasma PLP concentrations in unsupplemented children ranged from 58 to 78 nmol/L in the different groups, and the supplemented children had higher concentrations. This suggests that nonsupplemented intakes were considerably above an EAR level (if it could be determined), but it is not possible to set an EAR on the basis of these data.

Kirksey and colleagues (1978) studied the nutritional status of 12-to 14-year-old female adolescents. Thirteen percent of the subjects were judged to have inadequate B₆ status as assessed by erythrocyte alanine aminotransferase (α-EALT) ratios with average dietary B₆ intakes estimated to be 1.24 mg. Driskell and colleagues (1985) studying 12-, 14-, and 16-year-old adolescent girls reported 20 percent with marginal B₆ status and 13 percent with B₆ deficiency as indicated by abnormal α-EALT ratios. Daily dietary B₆ intakes in these groups averaged about 1.2 mg. In a follow-up longitudinal assessment of 12- and 14-year-old adolescent girls, Driskell and colleagues (1987) reported a prevalence of about 20 percent B₆ deficiency and almost 30 percent marginal deficiency with average dietary intakes of 1.25 mg B₆, again as assessed by α-EALT ratios. Although these data may suggest an EAR for adolescents somewhat below 1.2 mg, a number of concerns limit the usefulness of these data for deriving an EAR. As indicated previously, cutoff α-EALT ratios for B₆ deficiency have not been correlated with any clinical indicator of deficiency, and considerable variation in α-EALT ratios has been reported (Leklem, 1990). In addition, in the Driskell et al. (1985) study, the data were stratified by supplement intake ranging from less than 1 mg to greater than 3.5 mg. In each group about 60 percent of the subjects had adequate B₆ status as determined by α-EALT whereas the percentages of subjects with marginal and with deficient status were essentially identical. This seriously calls into question the appropriateness of using α-EALT ratios for assessing B₆ status.

In the absence of additional information, EARs and RDAs for these age groups have been estimated by using the method described in Chapter 2 and then extrapolating from adult values. For comparative purposes the requirement was also estimated by assuming that there is a direct relationship between protein intake and the B₆ requirement. The results of applying that method to data on protein intake from the Third National Health and Nutrition Examination Survey for children were judged to be unreasonably high. (See also the previous discussion about the protein-B₆ relationship.) As with adults, there is no compelling evidence to suggest that requirements within any of these age groups should be adjusted based on protein intake.

B₆ EAR and RDA Summary, Ages 1 through 18 Years

The RDA for B₆ 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 B₆; 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 B₆ the RDA is 120 percent of the EAR).

Men Ages 19 through 50 Years

Evidence Considered in Estimating the Average Requirement. The initial studies on B₆ requirements of young men are not very useful for determining an EAR because of the small number of repletion intakes. Most of these studies used the tryptophan load test to determine status. In a repletion study by Baker and colleagues (1964), eight young men on a diet containing 30 g/day of protein required 1.25 mg of PN (equivalent to 1.5 mg of food B₆) to return tryptophan metabolite excretion back to baseline after a 10-g tryptophan challenge dose, whereas subjects on a 100-g protein diet needed 1.5 mg of PN (equivalent to 1.9 mg of food B₆). This suggests an EAR considerably lower than 1.25 or 1.5 mg. Yess and coworkers (1964) placed six subjects on a diet containing 0.16 mg of PN and 100 g of protein for 55 days and investigated various tryptophan metabolites after a 2-g load. Supplementation with 0.6 or 0.9 mg of PN normalized tryptophan catabolites in nearly all the subjects, suggesting an EAR of less than 0.9 mg of PN (equivalent to less than 1.0 mg adjusted for food B₆).

Miller and Linkswiler (1967) measured tryptophan metabolites in 11 men receiving a basal diet containing 0.16 mg of PN and 54 or 150 g of protein. Over a 40-day period there was a slow increase in the excretion of tryptophan catabolites. The addition of 0.6 mg of B₆ corrected tryptophan catabolite excretion in most of the subjects, suggesting an EAR of less than 0.76 mg of PN (equivalent to less than 0.9 mg adjusted for food B₆). Complete normalization was observed with 1.5 mg (1.9 mg food B₆), which suggests that this level of intake exceeded the requirements of all subjects.

Linkswiler (1978) summarized a large number of early studies of men receiving various protein intakes. In subjects on a diet containing 100 g of protein, normalization of the load test required 1.0 to 1.5 mg, suggesting an EAR of considerably less than 1.0 to 1.5 mg. Miller and colleagues (1985) fed diets containing 1.6 mg of B₆ (from food and food fortified with PN) and 0.5, 1.0, or 2 g/kg of protein to eight young men (21 to 31 years of age). In this study, plasma PLP remained above 30 nmol/L, suggesting that 1.5 mg was substantially higher than the EAR.

In a study by Selhub and colleagues (1993), which adjusted for folate status, gender, and age, half of the subjects consuming freely selected diets containing an average of 1.3 mg of B₆ (as estimated from food frequency questionnaires and thus from food and vitamin supplements) had plasma homocysteine concentrations that were similar to those of subjects receiving much higher intakes of the vitamin. Subjects had not fasted or received a methionine load before being evaluated. Because no value for optimal homocysteine concentrations has been established and it is not clear that there is an advantage to driving homocysteine concentrations down to the lowest possible level, this suggests an EAR of less than 1.3 mg if plasma homocysteine is used as an index of adequacy.

It is difficult to derive a precise EAR for young men because most studies have identified a B₆ intake that restores test parameters to baseline values for all subjects, which may be considerably higher than that needed for health. Moreover, in most studies the levels of B₆ tested appear to have been in excess of an average requirement.

B₆ EAR and RDA Summary, Men Ages 19 through 50 Years. The various studies suggest an average requirement that ranges from less than 0.9 mg of food B₆ to considerably less than 1.9 mg. Taken together, the studies suggest an EAR of 1.1 mg/day.

The RDA for B₆ 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 B₆; 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 B₆ the RDA is 120 percent of the EAR).

Women Ages 19 through 50 Years

Evidence Considered in Estimating the Average Requirement. Six studies provide useful evidence about the amount of B₆ required by young women. Analysis of these studies focuses on the amount required for a plasma PLP value of at least 20 nmol/L. Brown and colleagues (1975) depleted 10 young women (average age 22) on a diet that contained 0.16 mg of B₆ and 78 g of protein for 28 days. The subjects were repleted with 1.0 or 2.2 mg of B₆ for 28 days. During the depletion period, plasma PLP concentrations fell from about 50 to about 14 nmol/L. On the two repletion regimens they increased to about 24 and 60 nmol/L. An EAR of just under 1.0 mg (equivalent to 1.2 mg based on food B₆) can be assessed based on a PLP concentration of ≥ 20 nmol/L.

Kretsch and coworkers (1995), using a depletion-repletion approach, studied a group of eight women (aged 21 to 30 years) on a high-protein diet (1.55 g/kg of body weight) and also investigated the effects of animal protein compared to plant protein. A formula diet plus 2 mg of PN was supplied for the 7-day adjustment period to develop baseline values. The depletion diet contained less than 0.05 mg of PN and was continued for 11 to 28 days (two subjects developed electroencephalogram [EEG] abnormalities). Repletion was with 0.5, 1, 1.5, or 2 mg/day of PN for 14 to 21 days. Between 1 and 1.5 mg of PN restored PLP to baseline, and 1.5 to 2 mg of PN restored xanthurenate to baseline values. In this study, plasma PLP in the baseline period averaged 25 nmol/L, a very low value compared with most studies, and dropped to less than 10 nmol/L during depletion. Surprisingly, plasma PLP fell slightly further during the first repletion with 0.5 mg of B₆. Values averaged close to 20 nmol/L on intake of 1 mg of B₆, suggesting an EAR of about 1 mg (1.25 mg of food B₆). Normalization of aminotransferase activation factors occurred with intakes of about 1.5 to 2 mg of B₆ and lagged behind the diet changes. Activation factor values increased during the depletion period and then increased further during the first repletion period, indicating that they had not stabilized within the depletion period. No differences in response were observed when comparing the animal to the plant protein source.

In a depletion-repletion study, Huang et al. (1998) studied B₆ requirements in young women (aged 28 to 34 years) fed a high-protein diet (1.55 g/kg of body weight). Subjects were fed a baseline diet (1.6 mg/day of B₆ from food) for 9 days and a depletion diet (0.45 mg/day also from food) for 27 days. Repletion diets of 1.26, 1.66, and 2.06 mg of B₆ were given for 21, 21, and 14 days, respectively, the additional B₆ provided in the form of pyridoxine hydrochloride once daily. As would be expected for a baseline intake of 1.6 mg of B₆, 1.5 to 2 mg of PN normalized plasma PLP and urinary 4-PA concentration to baseline values. No hematological or EEG abnormalities were observed during depletion, and the plasma PLP concentration was above 20 nmol/L at the end of the depletion period. With longer depletion, PLP values might have dropped more. At a B₆ intake of 1.26 mg, mean PLP was 38 nmol/L. On the basis of plasma PLP, an EAR should be between the two levels tested, 0.45 and 1.26 mg, probably less than 1 mg. If suggested cutoffs for erythrocyte aminotransferase concentrations were applied to ratios for α-EAST and α-EALT observed in the study, the EAR values would be around 1.3 mg based on α-EAST or above 1.7 mg based on α-EALT. Again, these values appear to lag with supplementation and may be overestimates.

Hansen and coworkers (1996b) investigated the effects of protein intake on B₆ status in women aged 19 to 38 years. The subjects received a diet containing 1.25 mg of PN, and protein intake was 0.5, 1, or 2 g/kg of body weight. At the two lower levels of protein intake, plasma PLP remained above 30 nmol/L. For women on the highest protein intake, PLP concentrations dropped below 30 nmol/L but remained above 20 nmol/L. There was a slight elevation in tryptophan catabolites after a load, which may have been due to the increased dietary tryptophan intake. The decrease in plasma PLP with increased protein intake was accompanied by a decrease in urinary 4-PA and urinary B₆, suggesting that more B₆ was retained by tissues under these conditions and that a steady state may not have been reached. This also calls into question whether the decrease in plasma PLP concentrations on a high-protein intake truly reflects a decrease in B₆ status. If a PLP cutoff of 20 nmol/L is applied, these studies suggest an EAR of considerably less than 1.25 mg.

Hansen and coworkers (1996a) also investigated the effects of diets low and high in PN glucoside on status indicators. Ten women (mean age 29 years) were fed two food-based diets each containing about 1.5 mg of B₆ of which 9 or 27 percent was as the naturally occurring B₆ glucoside. The status indicators demonstrated lower bioavailability of the glucoside, as expected, and suggested increased fecal excretion. On the low-glucoside diet, mean PLP concentrations were above 30 nmol/L, suggesting an EAR of less than 1.5 mg.

In a recent study, Hansen and coworkers (1997) investigated the B₆ requirements of young women receiving 85 g/day of protein. Subjects were adapted to a basal diet containing either 1.03 or 0.84 mg of B₆, which was fed for 15 days. The subjects then received three or four different levels of repletion for periods of 10 or 12 days. EAR values of less than 1.18 mg and 1.15 mg could be assessed based on plasma PLP concentrations. α-EAST ratios were in the normal range even in the basal period, which suggests an EAR of less than 0.8 mg. α-EALT was also normal in all subjects receiving the basal 0.84-mg diet, but half the subjects in the group receiving the basal 1.03-mg diet had values in the abnormal range. Although this suggests an EAR of 0.8 to 1.0 mg based on aminotransferase activation factors, these data are quite different from those reported in other studies.

Driskell and colleagues (1989) investigated the B₆ status of 15 obese black adult women (21 to 51 years of age). Dietary B₆ intake of the obese women and a nonobese control group averaged 1.18 mg, as assessed by dietary recall. Plasma PLP concentrations averaged 60 nmol/L in the obese group and 63 nmol/L in the control group, indicating no apparent effect of obesity on B₆ status. These data do, however, suggest an EAR of considerably less than 1.2 mg for adult women.

B₆ EAR and RDA Summary, Women Ages 19 through 50 Years. Based on the above evidence, the average requirement of PN is less than 1.0 mg. This is adjusted to 1.1 mg for food B₆.

The RDA for B₆ 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 B₆; 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 B₆ the RDA is 120 percent of the EAR).

Adults Ages 51 Years and Older

Evidence Considered in Estimating the Average Requirement. Studies of older individuals have been quite limited. Ribaya-Mercado and coworkers (1991) investigated B₆ requirements in 12 men and women over 60 years of age who received protein intakes of 0.8 or 1.2 g/kg of body weight. Depletion and repletion B₆ intakes were based on protein intake but averaged, respectively, about 0.15, 1.2, 1.8, and 2.6 mg in men and 0.1, 0.9, 1.3, and 1.9 mg in women. The source of B₆ during the repletion diets was a mixture of food and pyridoxine supplements. Individuals were kept on self-selected diets for baseline studies; these diets averaged 1.6 mg of B₆ for men and 1.4 mg for women. About 1.9 mg of B₆ was required to reach baseline values of plasma PLP and urinary 4-PA in both men and women consuming 1.2 g protein/kg of body weight. Normalization of erythrocyte aminotransferase activation factors required higher levels of B₆, but these activation factors did not stabilize during each period. The EAR based on a plasma PLP of at least 20 nmol/L would be less than 1.3 mg (less than 1.6 mg for food B₆) in men and about 0.9 mg (1.1 mg for food B₆) in women. Xanthurenate excretion in nearly half the men and all the women was near the baseline value at the equivalent of a food B₆ intake of 1.6 mg/day. The four subjects on the lower protein diet (0.8 g/kg of body weight) reached baseline values of plasma PLP at a lower level, 1.33 mg/day of B₆, yet xanthurenate excretion did not return to baseline for some of the subjects by the end of the three repletion periods.

Selhub and colleagues (1993) investigated the relationship among plasma homocysteine, dietary B₆, and plasma PLP concentration in subjects from the Framingham Heart Study aged 67 to 96 years. The data were adjusted for age, gender, and folate and B₁₂ intakes. No clear cutoff for homocysteine has been established, but about half the subjects whose B₆ intakes were 1.4 mg (and who had plasma PLP concentrations of 25 nmol/L) were in the same homocysteine range as subjects consuming much higher intakes. At a plasma PLP concentration of 20 nmol/L, homocysteine concentrations averaged about 13 μmol/L, and this corresponded to a dietary intake of about 1.3 mg B₆.

Meydani and coworkers (1991) reported impairments in interleukin-2 and lymphocyte proliferation in eight healthy elderly subjects (four men and four women) placed on a B₆-deficient diet of 0.3 μg/kg body weight/day for up to 20 days (approximately 0.17 mg/day for the men and 0.10 mg/day for the women). Restoration of various parameters of cell-mediated immunity to baseline values (obtained when the subjects had been on self-selected diets) required more than 22.5 μg/kg body weight/day, the second highest repletion level tested, because most of the parameters measured returned to baseline values at the highest repletion level tested in both the men and women, which was 33.75 μg/kg body weight/day or about 2 mg/day of B₆ for the women and 2.88 mg/day for the men. Some measures of immunity did not return to baseline during the study period whereas others exceeded baseline values only at very high (50 mg) B₆ intakes. No indication was given as to which level of any of these parameters indicated a dysfunction in immune response. Although the changes in these parameters with B₆ intake suggest they may be of potential use as indicators of B₆ status, it is presently not possible to determine a requirement from these data.

B₆ EAR and RDA Summary, Ages 51 Years and Older. From the above studies, it appears that the EAR for B₆ is higher than the EAR for younger men and women and, unlike that for younger adults, is higher for men than for women. The increase due to age and gender appears to be approximately 0.2 to 0.3 mg of food B₆/day.

The RDA for B₆ 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 B₆; 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 B₆ the RDA is 120 percent of the EAR).

Special Considerations

Note that the EARs (and RDAs) for adults were derived by using biochemical indicator cutoff values that have not been directly linked to clinical or physiological insufficiency. As described above, clinical symptoms of B₆ deficiency have been observed only in controlled studies during depletion with very low levels of B₆ and have never been seen at intakes of 0.5 mg/day or greater. This suggests that 1 mg/day is sufficient for most adults, but requirements may possibly be higher in individuals on very-high-protein diets.

Evidence Considered in Estimating the Average Requirement

Concentrations of indicators of B₆ status in plasma and blood decrease throughout pregnancy, especially in the third trimester (Cleary et al., 1975; Hamfelt and Tuvemo, 1972; Shane and Contractor, 1980). Many studies have demonstrated a drop in plasma PLP to about 10 nmol/L; this is substantially more than can be accounted for by increased blood volume. Fetal blood PLP concentrations in the second and third trimester and at term are much higher than in the mother, and significant fetal sequestration of the vitamin has been demonstrated by cord vein and artery differences (Cleary et al., 1975; Contractor and Shane, 1970; Shane and Contractor, 1980). Barnard and colleagues (1987) reported that much of the decrease in plasma PLP concentration during pregnancy is offset by increases in plasma PL, but other investigators have not observed this compensation (Contractor and Shane, 1970). Studies in animals suggest direct transport of PLP to the fetus (Contractor and Shane, 1971). In the pregnant rat about 15 percent of an intraperitoneal dose of PN is initially taken up by the uterus, placenta, and fetus.

Maintenance of plasma PLP concentrations at nonpregnant values requires about 2 mg/day of supplemental PN in the first trimester and between 4 and 10 mg/day in the third trimester (Cleary et al., 1975; Hamfelt and Tuvemo, 1972; Lumeng et al., 1976). Maintenance of other status indicators such as tryptophan metabolites at nonpregnant levels after a tryptophan load requires even higher intakes, but this test may be affected by hormonal changes (Shane and Contractor, 1980). It is not clear whether these changes in status indicators during pregnancy reflect poorer vitamin status or represent normal physiological changes during pregnancy. The latter is more reasonable. There is no a priori reason to use laboratory values for nonpregnant women as controls for pregnant women. There is also no evidence of significant problems in B₆ status during pregnancy despite the reduced levels of status markers. Schuster and coworkers (1981) examined the relationship between α-EAST ratios and Apgar scores in low-income mothers receiving an average of 1.3 mg/day of B₆ in the diet. A small but significant effect of B₆ status was noted. However, the possibility could not be eliminated that other variables unrelated to B₆ may have been responsible.

For an assumed body store of 1,000 μmol and a fetal, uterine, and placental accumulation of 15 percent, the fetus and placenta would accumulate approximately 25 mg of B₆. This would be about 0.1 mg/day averaged over gestation. With additional allowances made for the increased metabolic needs and weight of the mother and about 75 percent bioavailability of food B₆, an additional average need in pregnancy of 0.25 mg can be estimated. This increased need would be concentrated more in the second half of gestation. Unlike nutrients such as iron, B₆ is not stored in the body to any substantial extent, so it is unlikely that a surplus in early gestation would satisfy the increased need in the latter stages of gestation. Consequently, an extra 0.5 mg/day of B₆ can be reasonably justified to meet the need in the third trimester.

B₆ EAR and RDA Summary, Pregnancy

Although 0.5 mg/day of B₆ may overestimate the additional need in early gestation, it was considered judicious to err on the side of ensuring sufficiency and add 0.5 mg/day to the EAR for nonpregnant women throughout pregnancy. Because of the approximation involved, no additional adjustment for adolescence is included.

The RDA for B₆ 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 B₆; 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 B₆ the RDA is 120 percent of the EAR).

Evidence Considered in Estimating the Average Requirement

As described above, the B₆ concentration in human milk varies depending on the mother's B₆ intake, and some women consuming less than 2.5 mg/day of B₆ produce milk with a B₆ content that is not much higher than that associated with consuming formula that resulted in convulsions in infants because of low levels of B₆. There is some variation in reported human milk B₆ content, which may be due to methodological differences. Between individuals there are variations in milk B₆ content at similar B₆ intakes. Existing data previously described suggest that the amount of B₆ required to increase the milk B₆ content by a small increment is much higher than that increment and that the additional requirement for lactation is considerably in excess of that suggested by the amount secreted via lactation (Borschel et al., 1986; West and Kirksey, 1976). To ensure a milk B₆ concentration of 0.13 mg/L, it is estimated that about five times that amount of B₆ must be consumed in addition to the EAR of 1.1 mg/day.

B₆ EAR and RDA Summary, Lactation

In light of the evidence that low maternal intakes could lead to compromised B₆ status in the infant (Borschel, 1995), it would be prudent to add 0.6 mg of B₆ to the EAR of 1.1 mg for nonpregnant women, giving an EAR for lactation of 1.7 mg/day of B₆. Because this is an approximation based on a number of assumptions, no adjustment is made for adolescent females who are lactating.

The RDA for B₆ 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 B₆; 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 B₆ the RDA is 120 percent of the EAR).

Adverse Effects

No adverse effects have been associated with high intake of vitamin B₆ from food sources. This review focuses on pyridoxine, the form of B₆ that was consumed in the reports cited below. Large oral supplemental doses of pyridoxine used to treat many conditions, including carpal tunnel syndrome and premenstrual syndrome, have been associated with the development of sensory neuropathy and dermatological lesions (Cohen and Bendich, 1986; Schaumburg and Berger, 1988). The causal association between high-dose pyridoxine and neuropathy has been well documented in animals since 1940 (Unna and Antopol, 1940) and in humans since 1983 (Schaumburg et al., 1983).

Sensory Neuropathy. The first clinical report of pyridoxine-induced neurotoxicity in humans (Schaumburg et al., 1983) describes seven patients (five women and two men) with severe sensory neuropathy of the extremities after 2,000 to 6,000 mg/day of pyridoxine for 2 to 40 months. Four individuals were unable to walk. Neurological signs and symptoms were diagnosed through objective neurological assessment and improved in all patients after withdrawal of medication. Other reports of peripheral sensory neuropathy associated with high-dose pyridoxine therapy (1 to 4 g/day) appeared in the 1980s (Baer, 1984; Bredesen and Parry, 1984; De Zegher et al., 1985; Friedman et al., 1986). The pathogenesis of pyridoxine-induced peripheral sensory neuropathy and dose-response relationships have been well-described in animal models (Phillips et al., 1978; Schaeppi and Krinke, 1985). Review of the limited data involving lower pyridoxine doses (Bernstein and Lobitz, 1988; Del Tredici et al., 1985) reveals that the risk of developing sensory neuropathy decreases quite rapidly at doses below 1 g/day.

Other Adverse Effects. Painful and disfiguring dermatological lesions were reported in humans after consumption of 2 to 4 g/day of pyridoxine for more than 1 year (Friedman et al., 1986; Schaumburg and Berger, 1988). However, the limited data fail to demonstrate a relationship between this endpoint and the dose or duration of treatment. Also, the mechanism of pyridoxine dermatoses remains unclear (Schaumburg and Berger, 1988).

There are isolated, uncontrolled case reports of congenital defects (Donaldson and Bury, 1982; Gardner et al., 1985; Philpot et al., 1995) and B₆ dependency (Hunt et al., 1954) in newborns whose mothers were treated with pyridoxine during the first half of pregnancy. No controlled data in humans were found that confirm these findings. Observational data in women taking up to 200 mg/day of pyridoxine orally (Ellis, 1987; Weinstein et al., 1944) and up to 100 mg/day of pyridoxine parenterally (Dorsey, 1949; Hart and McConnell, 1943; Weinstein et al., 1944; Willis et al., 1942) during the first trimester of pregnancy report no adverse effects. Controlled studies in animals (Hendrickx et al., 1985; Khera, 1975; Schumacher et al., 1965) show no evidence of teratogenicity.

Very high pyridoxine doses (200 to 600 mg/day parenterally; 600 mg/day orally) after delivery have been used to successfully inhibit or suppress lactation (Foukas, 1973; Marcus, 1975; Scaglione and Vecchione, 1982). However, the results of other clinical trials on the antilactogenic effect of pyridoxine are conflicting (Scaglione and Vecchione, 1982). No evidence of antilactogenic effects at lower oral doses was found.

Summary

On the basis of considerations of causality, relevance, and the quality and completeness of the database, sensory neuropathy was selected as the critical endpoint on which to base a Tolerable Upper Intake Level (UL). The data fail to demonstrate a causal association between pyridoxine intake and other endpoints reviewed (e.g., dermatological lesions and B₆ dependency in newborns).

Adults

Data Selection. Studies involving long-term oral administration of pyridoxine at doses of less than 1 g/day (Berger and Schaumburg, 1984; Bernstein and Lobitz, 1988; Dalton, 1985; Dalton and Dalton, 1987; Del Tredici et al., 1985; Parry and Bredesen, 1985) were considered more relevant for deriving a UL than studies involving short-term or single-dose intake, parenteral administration, or supplemental doses greater than 1 g/day.

Identification of a No-Observed-Adverse-Effect Level (NOAEL) and a Lowest-Observed-Adverse-Effect Level (LOAEL). A NOAEL of 200 mg/day can be identified by a critical evaluation of two studies (Bernstein and Lobitz, 1988; Del Tredici et al., 1985). Bernstein and Lobitz (1988) treated 70 patients with diabetic neuropathy or carpal tunnel syndrome with 100 to 150 mg/day of pyridoxine—some for up to 5 years. Despite rigorous neurological examination and testing, sensory neuropathy was not detected in any of these patients. Similarly, Del Tredici and colleagues (1985) reported on 24 patients treated for carpal tunnel syndrome with pyridoxine at doses of 100 to 300 mg/day for 4 months. A NOAEL of 200 mg/day represents the average of 100 and 300 mg/day.

Additional investigations report no neuropathy in hundreds of individuals given pyridoxine doses of 40 to 500 mg/day (Brush et al., 1988; Ellis et al., 1979; Mitwalli et al., 1984; Tolis et al., 1977). Although these studies were not as carefully executed or reported as the study by Bernstein and Lobitz (1988), they provide supportive evidence for a NOAEL of 200 mg/day.

A LOAEL of 500 mg/day was identified from a case report by Berger and Schaumburg (1984). These investigators reported that a 34-year-old woman who had been taking 200 mg/day of pyridoxine for 2 years without adverse effects developed sensory neuropathy when she increased her daily dose to 500 mg supplemented occasionally with a 300-mg dose once a week. Another report by Schaumburg and coworkers (1983) showed severe sensory neuropathy in seven adults after pyridoxine intakes that started at 50 to 100 mg/day and were steadily increased to 2 to 6 g/day over 2 to 40 months. None of these patients experienced signs or symptoms at doses of less than 2 g/day.

Several reports show sensory neuropathy at doses lower than 500 mg/day. Tabulated data in the report by Parry and Bredesen (1985) show that two patients who took 500 mg/day of pyridoxine for 8 and 36 months developed sensory neuropathy and as did one patient who took 100 to 200 mg/day for 36 months. This finding conflicts with the weight of evidence showing that daily doses in the range of 100 to 200 mg are not associated with the development of this condition. It is not clear from the report whether the neurological symptoms of the patient taking 100 to 200 mg/day were confirmed by a clinical neurological examination. The report notes that clinical neurological examinations were performed on about half of the patients. In addition, the report notes that electrophysiological studies were performed on seven patients and sural nerve biopsies were performed on two patients. It is not clear from the report which patients were examined clinically and which received additional tests. A local television report publicizing this syndrome before the study may have biased the selection of patients and reporting of neurological symptoms.

Two additional studies that report sensory neuropathy at doses of less than 200 mg/day (Dalton, 1985; Dalton and Dalton, 1987) warrant examination. In a letter to the editor, Dalton (1985) reported sensory neuropathy (characterized as burning, shooting, and tingling pains; paresthesia of limbs; clumsiness, ataxia; or perioral numbness) in 23 of 58 women (40 percent) being treated with 50 to 300 mg/day of pyridoxine for premenstrual syndrome. This case report contains methodological flaws including lack of information on the duration of treatment, use of other medications or herbal preparations, and lack of confirmatory information on actual doses consumed.

In a subsequent publication, which may have included patients from the earlier case report (Dalton, 1985), Dalton and Dalton (1987) retrospectively studied 172 women who were attending a private practice specializing in premenstrual syndrome and who were reported to have taken 50 to 500 mg/day of pyridoxine. The subjects were divided into two groups: 103 women who complained of neurological symptoms and 69 who did not. Neurological symptoms were not adequately detailed in this study. In addition, the actual doses may be underestimated because the patients were also taking vitamin supplements. In summary, the weaknesses of this study and the inconsistency of the results with the weight of evidence pertaining to the safety of higher doses of pyridoxine rule out the use of these data to determine a UL.

Uncertainty Assessment. An uncertainty factor (UF) of 2 was selected based on the limitations of the data involving pyridoxine doses of less than 500 mg/day.

Derivation of a UL. The NOAEL of 200 mg/day was divided by the UF of 2 to obtain a UL of 100 mg/day for adults:

B₆ UL Summary, Ages 19 Years and Older

Other Life Stage Groups

Some concern for pregnant and lactating women could arise from the data available in the literature on congenital defects, B₆ dependency, and antilactogenic effects (Donaldson and Bury, 1982; Foukas, 1973; Gardner et al., 1985; Hunt et al., 1954; Marcus, 1975; Philpot et al., 1995; Scaglione and Vecchione, 1982). Scientifically based, controlled studies designed to assess the potential adverse effects of pyridoxine intake by pregnant and lactating women are lacking. As noted above, the weight of the evidence from controlled studies in animals during pregnancy reveals no adverse effects related to teratogenicity, and the evidence from humans reveals no adverse effects from intakes up to 200 mg/day. Therefore, a UL of 100 mg/day was set for pregnant and lactating women as well. The ULs for children and adolescents were calculated from the UL for adults by using the method described in Chapter 3; this method adjusts for body size.

B₆ UL Summary, Ages 1 through 18 Years, Pregnancy, Lactation

Special Considerations

A review of the literature failed to identify special subgroups that are distinctly susceptible to sensory neuropathy after excess pyridoxine intake.