20b.1 Riboflavin

Riboflavin (7,8-dimethyl-10-ribityl-isoalloxazine) was first synthesized in 1935. Its structure consists of an isoalloxazine ring attached to a ribityl side chain Figure 20b.1. A detailed review of ribo­flavin and health is provided elsewhere (Powers, 2003; Thakur et al., 2017; Suwannasom et al., 2020).

20b.1.1 Functions of ribo­flavin

Ribo­flavin is a component of flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD) which act as coenzymes of different flavo­proteins involved in oxidation-reduction reactions that are central to energy production, intermediary metab­olism, drug metab­olism as well as the maintenance of anti­oxidant status (Figure 20b.1).

Of the two coenzymes, FMN is formed first, from free ribo­flavin, by ATP-dependent phos­phoryla­tion; a reaction catalyzed by cytosolic flavo­kinase (EC 2.7.1.26). Next, most of this FMN is combined with a molecule of ATP to form FAD; this step is catalyzed by the FAD-dependent FAD syn­thetase (EC 2.7.7.2). In turn, FAD can be converted into forms covalently bound to tissue proteins. Thyroid hormones regulate the synthesis of FMN and FAD as well as the formation of the covalently bound flavins (Rivlin, 1970; Pinto and Rivlin, 1979).

Both FMN and FAD are also cofactors for several enzymes, including glutathione reductase, xanthine oxidase, L‑amino oxidase, and nico­tinamide adenine dinucleo­tide (NAD) dehy­drogenase. Results from animal studies also suggest that flavins are cofactors in the cyclical β‑oxi­dation of fatty acids (Olpin and Bates, 1982a; Olpin and Bates, 1982b). Around 84% of human flavo­proteins are FAD‑depen­dent and only 16% use FMN as a cofactor (Leinhart et al., 2013).

Ribo­flavin coenzymes are also involved in the metab­olism of four other vitamins — folate, vitamin B₁₂, vitamin B₆, and niacin. Flavin adenine dinucleo­tide is a cofactor for 5,10‑methyl­enetetra­hydro­folate reductase (MTHFR; EC 1.7.99.5), a key enzyme in folate metab­olism involved in the conversion of 5,10‑methyl­enetetra­hydro­folate to 5-methyl­tetra­hydro­folate, which is required for the remethylation of homo­cysteine to meth­ionine. Both FMN and FAD act as cofactors for the enzyme meth­ionine synthase reductase (EC 1.16.1.8) which is responsible for the regeneration of methyl­cobalamin, the bio­logically active form of vitamin B₁₂ that is also involved in the remethyl­ation of homo­cysteine to meth­ionine. In addition, FMN acts as a cofactor for pyri­dox­ine (pyridox­amine) phos­phate oxidase (PPO; EC 1.4.3.5), an important enzyme in vitamin B₆ metab­olism that converts the 5‑phos­phates of both pyri­dox­ine and pyri­dox­amine to the coenzyme pyri­dox­al‑5´‑phos­phate. Finally, the FAD‑dependent kynurenine hydroxylase is involved in the conversion of tryptophan to niacin. As a result, severe ribo­flavin deficiency can cause functional deficiency with disturb­ances in the meta­bolic pathways of these vitamins.

Absorption of ribo­flavin occurs mainly in the proximal small intestine by a saturable carrier-mediated mechanism; the maximal amount of ribo­flavin that can be absorbed from a single dose is about 27mg (Zempleni et al., 1996). Intestinal micro­flora is able to synthesize ribo­flavin, some of which is absorbed in the colon by a carrier-mediated process (Said et al., 2000; Said, 2013 ). The vitamin is transported in plasma as free ribo­flavin complexed with albumin and other proteins, mainly immuno­globulins, which also bind flavin coenzymes (Innis et al., 1985).

20b.1.2 Deficiency of ribo­flavin in humans

The classical signs of ribo­flavin deficiency, termed aribo­flavinosis, are angular stomatitis, cheilosis, glossitis and anemia. Corneal vascular­ization, dermato­logical changes, and neuro­logical alterations may also occur but are not specific for aribo­flavinosis (Northrop-Clewes and Thurnham, 2012). Some environ­mental factors may also influence the clinical signs and symptoms of ribo­flavin deficiency.

Aribo­flavinosis is rarely encountered in isolation, usually occurring in association with other vitamin deficiency states. In a study of Irish elderly, 49% had suboptimal ribo­flavin status, 39% had suboptimal vitamin B₆ status, and 21% had con­current ribo­flavin and vitamin B₆ deficiencies. After supple­mentation with ribo­flavin alone, both the ribo­flavin deficiency and the low plasma pyri­dox­al‑5´‑phos­phate levels were corrected (Madigan et al., 1998).

Some evidence suggests that ribo­flavin deficiency is linked with impaired mobilization of iron from intra­cellular stores (i.e. hepatic ferritin), increased rate of iron loss from the gastro­intestinal tract, as well as decreased absorption of iron (Powers, 2003). This could explain the low hemoglobin con­cen­trations and hypo­chromic anemia which often are observed in ribo­flavin deficiency (Aljaadi et al., 2019). Studies conducted in population groups with a com­promised ribo­flavin status such as school children (Buzina et al., 1979; Charoenlarp et al., 1980), men (Fairweather-Tait et al., 1992), pregnant (Decker et al., 1977; Suprapto et al., 2002; Ma et al., 2008), lactating (Powers et al., 1985) and women of reproductive age (Powers et al., 2011), have shown that ribo­flavin supple­mentation leads to improve­ment of the hemat­ological status.

Ribo­flavin deficiency has been described in under­nourished pop­ulations in several low income countries, notably among women and children in The Gambia (Bates et al., 1981, 1994) , some elderly persons in Guatemala (Boisvert et al., 1993a), children in Côte d'Ivoire (Rohner et al., 2007), and in some adol­escent refugees from Bhutan living in south­eastern Nepal (Blanck et al., 2002). Infants of mothers with a low ribo­flavin status during gestation are also likely to be born ribo­flavin-deficient (Bates et al., 1982). An accumulating body of evidence based on small as well as pop­ulation-based surveys from developed countries has shown that a sub­optimal ribo­flavin status is more common than previously recognized and those pre­domin­antly affected are adol­escents and young women (Bates et al., 2016; Ward et al., 2020; Aljaadi et al., 2019; Jungert et al., 2020).

Several conditions, including alcoholism, diabetes mellitus, liver disease, thyroid and adrenal insufficiency, and gastro­intestinal and biliary obstruction, may precipitate or exacerbate ribo­flavin deficiency (Rivlin, 2007). Alcohol causes deficiency by interfering with both the digestion and the intestinal absorption of ribo­flavin S (Pinto et al., 1987; Subramanian et al., 2013 ). Anti­epileptic and psycho­tropic drugs such as chlor­promazine, imipramine, and amitriptyline (Apeland et al., 2003; Pinto et al., 1982), as well as some anti­malarial drugs such as quinacrine (Dutta et al., 1985) all inhibit the conversion of ribo­flavin to its active coenzyme derivatives. Drugs, such as tetra­cycline, theo­phylline, and caffeine, as well as metals, such as zinc, copper, and iron, may chelate or form complexes with ribo­flavin and, hence, affect its bio­avail­ability (Sauberlich, 1985).

More vulnerable to ribo­flavin deficiency are individuals with 677C→T poly­morphism in the gene encoding MTHFR enzyme because the variant enzyme has an impaired activity as a result of reduced affinity to the cofactor FAD (Yamada et al., 2001); this typically leads to elevated homo­cysteine con­cen­trations (Frosst et al., 1995). Marked lowering of homo­cysteine has been achieved in people with the impaired variant (TT genotype) with a low dose of ribo­flavin (McNulty et al., 2006). Moreover, genome-wide association studies (Ehret et al., 2011) and clinical studies (Qian et al., 2007; Yang et al., 2014) provide evidence linking the MTHFR 677C→T poly­morphism with blood pressure and increased risk of hyper­tension and hyper­tension in pregnancy by up to 87%. Importantly, emerging evidence from random­ized trials highlights that ribo­flavin supple­mentation can lower blood pressure specifically in adults with TT genotype (Horigan et al., 2010; Wilson et al., 2013). This novel role of ribo­flavin could have important public health implications, considering that the frequency of the variant TT genotype is 10–12% worldwide but it could reach up to 32% in some countries (e.g., Mexico; Wilcken et al., 2003).

Ribo­flavin deficiency has been implicated as a risk factor for cancer in animal studies (Powers, 2003), but the epidemiological evidence based on several meta-analyses is less consistent (Yoon et al., 2016; Yu et al., 2017; Ben et al., 2019; Zeng et al., 2020).

Studies have shown that in malaria endemic regions, individuals with ribo­flavin deficiency are relatively resistant to this infection and have a lower level of parasitemia, but the course of the disease may be more severe than in people with adequate ribo­flavin status (Das et al., 1988).

20b.1.3 Food sources and dietary intakes

Ribo­flavin in foods is present predominantly in the form of FAD and only small quantities are available as FMN. Some flavins bound covalently to protein are also found in certain foods but are largely unavailable as nutritional sources of ribo­flavin; only limited amounts apparently undergo digestion and absorption (Chia et al., 1978).

The major food sources of ribo­flavin are dairy products, especially milk, and meat and fish; most plants contain only small amounts of ribo­flavin. As a result, individuals consuming mainly plant-based diets may be at risk for ribo­flavin deficiency. Ribo­flavin enrichment of flour is mandated in the US and some other countries with the aim of restoring the losses of the vitamin during milling and refining processes. Most of the ready-to-eat breakfast cereals are fortified with ribo­flavin in developed countries. Although it is heat-stable, losses of ribo­flavin do occur if it is exposed to the light and, as it is water soluble, by leaching into the cooking water (Powers, 2003). Bioavailability of ribo­flavin from food is reported to be around 95% (IOM, 2000). There are limited data on the relative bio­availability of ribo­flavin from different food sources but a study using stable isotopes and kinetic modeling did not find a significant difference in ribo­flavin absorption from milk and spinach (Dainty et al., 2007).

Nationally representative surveys from the US, Australia, Ireland, and the UK showed that milk and dairy products, meat and ready-to-eat breakfast cereals are the main dietary contributors to ribo­flavin intake (IOM, 2000; Australian Bureau of Statistics, 2014; National Adult Nutrition Survey, 2011; Bates et al., 2016).

20b.1.4 Effects of high intakes of ribo­flavin

In humans, there is no evidence for ribo­flavin toxicity as a result of excessive intakes (EFSA, 2017). The absorption of orally administered ribo­flavin from both vitamin supple­ments and from natural foodstuffs is limited, and high intakes are rapidly excreted in the urine. Even when 400mg/d of ribo­flavin was given orally with meals for at least 3 months, no short-term side effects were reported (Schoenen et al., 1994). In view of the limited data on adverse effects from high intakes of ribo­flavin, no Tolerable Upper Intake Level for ribo­flavin was set by the U.S. Food and Nutrition Board (IOM, 2000).

20b.2 Biomarkers of ribo­flavin status

As the signs and symptoms of ribo­flavin deficiency are not very specific, diagnosis of a deficiency state is difficult when based exclusively on clinical assessment. Consequently, bio­chemical tests are essential for confirming clinical cases of ribo­flavin deficiency and for estab­lishing sub­clinical deficiencies. Several tests are available, and these are discussed in the following sections. A combination of bio­markers is generally preferred: erythrocyte glutathione reductase activity and urinary ribo­flavin excretion are the most used.

20b.2.1 Erythrocyte glutathione reductase activity

The measurement of the activity coefficient of erythrocyte glutathione reductase (EGR) (EC 1.6.4.1), an erythrocyte enzyme that depends on a cofactor derived from ribo­flavin, is the preferred method for assessing ribo­flavin status. It provides a measure of tissue saturation and long-term ribo­flavin status.

Glutathione reductase is a nicotinamide adenine dinucleotide phosphate (NADPH), a FAD-dependent enzyme, and the major flavo­protein in erythrocytes. It catalyzes the oxidative cleavage of the disulfide bond of oxidized glutathione (GSSG) to form reduced glutathione (GSH):

$${\text{GSSG + NADPH + H}}^{+}{\text{→ 2GSH + NADP}}^{+}$$ The activity of EGR is measured spectrophotometrically by monitoring the oxidation of NADPH to NADP⁺ at 340nm, with and without the presence of added FAD coenzyme. As an alternative, the production of GSH can be monitored colorimetrically. The test involves the following steps:

1. The basal activity of EGR is measured. This represents the endogenous enzyme activity and depends on the amount of FAD coenzyme in erythrocytes.
2. The activity of EGR with excess FAD coenzyme added in vitro is then deter­mined. This represents the maximum potential EGR activity and is referred to as “stimulated” activity.
3. The activity coefficient (EGR AC) is then derived by the ratio of the “stimulated” over the basal EGR activity which indicates the degree of unsaturation of the enzyme with the coenzyme:

   $$\text{EGR AC =}\frac{\text{activity (with added FAD)}}{\text{ basal activity (without added FAD)}}$$

The basal and stimulated EGR activities can be expressed per gram of hemoglobin, per number of erythrocytes, or in terms of the volume of erythrocytes (in mL).

The degree of in vitro stimulation of EGR activity depends on the FAD saturation of the apoenzyme, which, in turn, depends on the availability of ribo­flavin. If the vitamin status is normal, then the values of the basal and stimulated activity are similar, and the EGR activity coefficient is very close to 1.0. In persons with ribo­flavin deficiency, the basal EGR activity falls and the in vitro stimulation by FAD rises; thus, higher values of EGR activity coefficient are indicative of a lower ribo­flavin status. The result also can be presented as “percentage of stimulation” (PS) which is calculated from the EGR AC as follows: $$\text{PS = (EGR AC × 100) − 100}$$

Meta-analysis of human supple­mentation trials confirmed that the EGR AC is a useful, stable, and sensitive measure of ribo­flavin status, reflecting the degree of tissue saturation ranging from severe deficiency to normal status (Hoey et al., 2009). It is not very sensitive, however, for detecting changes of ribo­flavin status in replete individuals (Powers, 1999). In experimentally controlled human studies, concomitant increases in EGR AC in response to decreases in intakes of ribo­flavin have been reported (Figure 20b.2).

Nevertheless, the increase in the EGR AC does not continue indefinitely; ribo­flavin intakes below 0.5mg/day do not produce any further increases in EGR AC (Sterner and Price, 1973). Consequently, the extent to which the EGR AC is elevated does not necessarily indicate the degree of ribo­flavin deficiency. Hence, it is not surprising that consistent correlations between EGR AC values and clinical signs of ribo­flavin deficiency have not always been observed (Bates et al., 1981).

Measurement of EGR AC has been used to monitor ribo­flavin status only in a few population-based studies such as the UK National Dietary and Nutrition Surveys Rolling Programme which includes all age and sex population groups (Bates et al., 2016), Irish National Adult Nutrition Survey (Kehoe et al., 2018) and the older national surveys from Germany (Heseker et al., 1992) and the Netherlands (Löwik et al., 1994). However, the assay has been used more frequently in cohort studies of UK older adults (Bailey et al., 1997), Irish adults (Jungert et al., 2020) and older people (Madigan et al., 1998; Moore et al., 2019), Canadian women of reproductive age (Aljaadi et al., 2019) and older adults (Whitfield et al., 2019), Spanish adults (Mataix et al., 2003; García-Minguillán et al., 2014) as well as French adults (Hercberg et al., 1994). Ribo­flavin status of different population groups assessed by EGR AC has been reported in studies conducted in developing countries including some in Africa (Bates et al., 1981; 1994; Ajayi, 1984; 1985; Ajayi and James, 1984), Central America (Boisvert et al., 1993a), and Asia (Thurnham et al., 1982; Blanck et al., 2002; Whitfield et al., 2015; Aljaadi et al., 2019).

Comparison of the prevalence of deficient ribo­flavin status among these studies is often difficult because the method of EGR AC assessment has not been standardized and as shown in a systematic evaluation of the existing methods, variations in the laboratory protocols applied in various studies could produce significantly different results (Hill et al., 2009). Moreover, the cutoff values used to denote suboptimal status and deficiency vary considerably from one study to another. However, EGR AC results from different populations generated by an identical protocol in one laboratory and by using the same cutoff showed very comparable data for the prevalence of ribo­flavin deficiency in Canadian and Irish older adults (26% vs 25.5%; Whitfield et al., 2019; Jungert et al., 2020) and in Canadian women of reproductive age and younger Irish adults (40% vs  52%; Aljaadi et al., 2019; Jungert et al., 2020) whereas 80% and 71% of the investigated Malaysian and Cambodian women of reproductive age, respectively were reported to be ribo­flavin deficient (Aljaadi et al., 2019; Whitfield et al., 2015).

In some of these surveys, significant inverse relationships between the EGR AC and ribo­flavin intakes have been observed (Gregory et al., 1990; Hercberg et al., 1994; Bailey et al., 1997; Bates et al., 1999; Hoey et al., 2007); results from the UK survey of British adults are given in Table 20b.1.

Table 20b.1: Correlation coefficients between the erythrocyte glutathione reductase activity coefficient and dietary riboflavin intakes for U.K. adults. Data from Gregory et al., 1990 .

	Erythrocyte glutathione reductase activity coefficient
Dietary intake	Men	Women
Total riboflavin (incl. supple­ments)	−0.13 ( p < 0.01)	−0.24 ( p < 0.01)
Riboflavin from food sources	−0.23 ( p < 0.01)	−0.31 ( p < 0.01)

In the UK study of Norwich elderly (Bailey et al., 1997), initial EGR AC values for both males and females were significantly correlated with those measured 2y later (Table 20b.2), suggesting that EGR AC may be a reliable measure of long-term bio­chemical ribo­flavin status of individuals. This finding is consistent with earlier studies (Rutishauser et al., 1979).

There have been several attempts to establish a relationship in humans between a specific physiological function and EGR AC values, but the results have been inconclusive. Some reports of effects on work performance, neurovascular coordination, and iron handling with EGR AC values ≥ 1.7 have been reported (Prasad et al., 1990; Fairweather-Tait et al., 1992; Bates et al., 1994).

Table 20b.2: The correlation of EGR activity coefficients measured on 99 subjects initially and after 2y. Signif­icant correlation does not imply agreement between initial and follow-up mean values. Data from Bailey et al., British Journal of Nutrition 77: 225–242, 1997.

Elderly subjects	n	r	Significance
Males	37	0.411	< 0.02
Females	62	0.359	< 0.01

In some supple­mentation studies of subjects with subclinical ribo­flavin deficiency, no improvement in physical performance or endurance has been reported (Prasad et al., 1990; Winters et al., 1992; van der Beek et al., 1994). However, a supple­mental ribo­flavin dose of 4mg/day given for 8wks appeared to have a beneficial effect on hemat­ologic status in UK young women with a low ribo­flavin status (EGR AC > 1.65) at baseline (Powers et al., 2011). Furthermore, there is evidence that a low dose ribo­flavin supple­mentation can lower an elevated blood pressure, specifically in adults who are genetically predisposed to develop hypertension (Horigan et al., 2010; Wilson et al., 2013; see Section 20b.1.2 on deficiency of ribo­flavin in humans).

Some epidemiological studies also have linked EGR AC values with various health effects. In a study of adolescent Bhutanese refugees living in Nepal, for example, in whom the prevalence of angular stomatitis was 27%, those with angular stomatitis had significantly higher EGR AC values than those without (i.e., 2.2 ± 0.4 vs. 2.0 ± 0.3; p = 0.02). The adjusted odds ratio for angular stomatitis and low ribo­flavin status was 5.1 (95% CI: 1.55, 16.5) (Blanck et al., 2002).

In addition, a large cross-sectional study of Irish older adults showed that ribo­flavin deficiency (EGR AC > 1.46) was independently associated with an increased risk of depression after adjustment for covariates (OR 1.56, 95% CI: 1.10, 2.00,  p=0.012) (Moore et al., 2019).

          Factors affecting EGR AC values

FAD concen­trations used in the assay to stimulate EGR can affect the EGR AC values obtained. Concen­trations of FAD > 5µmol result in lower normal ranges of EGR ACs, in comparison with FAD concen­trations ranging from 1–3µmol (Garry and Owen, 1976; Rutishauser et al., 1979; Hill et al., 2009).

The length of the pre-incubation of reagents with EGR enzyme appears also to be critical. A short pre-incubation period could underestimate EGR AC values because less time is available for FAD to bind to the enzyme in the hemolysate (Thurnham and Rathakette, 1982). Hill et al. (2009) found that 30min would be the optimal length of the pre-incubation.

Age of erythrocytes may also influence the EGR activity (Powers and Thurnham, 1981) as concen­trations are declining with the aging of the cells.

Age of the subjects may affect the EGR activity. In some early studies a trend toward higher EGR ACs with increasing age, irrespective of sex, has been observed (Garry et al., 1982; Wright et al., 1995). However, more recent investigations including the UK population-based survey, have reported consistently higher EGR AC values in younger compared with older adults (Bates et al., 2016; García-Minguillán et al., 2014; Jungert et al., 2020). It is unknown whether this difference in ribo­flavin status with age can be explained only by variations in dietary ribo­flavin intake.

Genetic disturbances, including glucose‑6-phosphate dehydrog­enase deficiency and hetero­zygous β‑thalassemia, are associated with disorders in erythrocyte flavin metab­olism (Prentice et al., 1981; Anderson et al., 1987; 1993), which can result in misleading EGR AC test results. For example, in cases of glucose‑6-phosphate dehy­drogenase deficiency, there is increased avidity of the EGR for FAD, resulting in EGR AC values within the normal range, even in the presence of clinical signs of ribo­flavin deficiency (Thurnham, 1972). Glucose‑6-phosphate dehy­drogenase deficiency is one of the most common enzyme defects in humans and it is highly prevalent among populations in malaria-endemic areas of sub-Saharan Africa and the Middle East where approximately up to 7.5% of people are affected (Nkhoma et al., 2009). It is estimated that glucose‑6-phosphate dehy­drogenase deficiency occurs in around 10% of Americans of African descent (Frischer et al., 1973). In contrast, in hetero­zygous β‑thalassemia, there is an inherited slow red-cell metab­olism of ribo­flavin to FMN and FAD and a high stimulation of the erythrocyte glutathionine reductase by extraneous FAD (Anderson et al., 1993).

Pyri­dox­ine deficiency also interferes with the EGR AC test, resulting in a decreased erythrocyte glutathione reductase activity but no change in the activity coefficient, probably arising from a decrease in apoenzyme. No comparable effects have been observed for other vitamin deficiencies such as thiamin and vitamin C.

Disease states, including iron-deficiency anemia (Ramachandran and Iyer, 1974), severe uremia, cirrhosis of the liver, and hypothyroidism lead to increased erythrocyte glutathione reductase activity.

Conditions of negative nitrogen balance lead to a fall in EGR AC values. In pregnant women in The Gambia, for example, EGR AC values fell in association with a decline in body weight during the rainy season, despite evidence that malnutrition had actually increased (Bates et al., 1981). Similar findings have been reported in preschool children with upper respiratory tract infections and measles (Bamji et al., 1987).

          Interpretive criteria

There is still uncertainty regarding the most appropriate threshold value indicative of normal ribo­flavin status. An EGR AC value of < 1.3 is generally considered to represent saturation of the tissues with ribo­flavin. This value was derived from the ribo­flavin depletion-repletion study in young men (Figure 20b.2) as well as from the supple­mentation study in elderly people with ribo­flavin deficiency (Boisvert et al., 1993b) where urinary excretion of ribo­flavin started to increase sharply (a point reflecting body saturation) at EGR AC below 1.3. A systematic review including 18 supple­mentation studies concluded that a cut-off of 1.3 should be considered as the “upper limit of a normal range” (Hoey et al., 2009). Sadowski 1992), however, used an upper limit of 1.34 in a study of apparently healthy elderly. This value was based on the mean plus 2 SD of the EGR AC value derived from a large sample of Boston elderly aged ≥ 60y. Cutoff EGR AC values of 1.3–1.4 for suboptimal and > 1.4 for deficiency states are often used (Wilson et al., 2013; Whitfield et al., 2015; Aljaadi et al., 2019; Jungert et al., 2020); these cutoffs appear to be independent of sex. Generally, EGR AC values in the range from 1.5 to 2.5 were found in population groups consuming very low intakes of ribo­flavin, wih evidence of clinical ribo­flavin deficiency (Bates et al., 1981; Thurnham et al., 1982; Lo, 1985). In a study of adolescent Bhutanese refugees, Blanck et al. (2002) used an EGR AC cutoff of > 1.7, and based on this value, 86% had low ribo­flavin status.

Considering that various methods for EGR AC assessment generate slightly different results (Hill et al., 2009), it is reasonable to expect that the cutoff value would be dependent on the analytical method used; thus, reference ranges developed in the local laboratories should be considered.

Table 20b.3: Median and 97.5^th percentile values for the activity coefficient (AC) of erythrocyte glutathione reductase (EGR) in UK adults by age and sex. Data from Bates et al. (2016).

	EGR activity coefficient
	19–64y	≥ 65y
Men Median	1.31	1.26
97.5th percentile	1.99	1.61
Women Median	1.33	1.28
97.5th percentile	1.94	1.71

Information on the EGR activity coefficients from the UK National Dietary Nutrition Survey Rolling Programme, which includes data for all age groups split by sex is available (Bates et al, 2016). The median and upper 2.5^th percentile of the adult population by age and sex are presented in Table 20b.3

          Measurement of erythrocyte glutathionine reductase activity

Only small samples of blood are required for the EGR assay, and fasting samples are not necessary (Komindr and Nichoalds, 1980). Either EDTA or heparin can be used as an anti­coagulant. Erythro­cytes must be washed and lysed, and the assay must be performed immediately after blood samples are drawn because FAD is a labile compound (Shenkin and Roberts, 2016). Alternatively, the hemolyzed samples can be kept frozen for over a year at −70°C without loss of EGR activity.

Generally, glutathione reductase activity is measured using an enzyme-coupled kinetic assay, although some colorimetric methods are also available (Sauberlich, 1984). An automated method involving the use of a centrifugal analyzer has been developed (Mak and Swaminathan, 1988; La Rue et al., 1997). An alternative procedure, using well plates and a plate reader, has also been described (Sauberlich, 1999). Very small volumes of whole blood instead of erythrocytes can be used for the measurement of glutathionine reductase activity, but the method, although simpler, is less sensitive. Care must be taken to ensure that the protocols for the assay are carefully specified and followed so that results can be compared among studies. As noted previously, the EGR assay is invalid for persons with glucose‑6- phosphate dehy­drogen­ase deficiency.

20b.2.2 Urinary ribo­flavin excretion

Flavins are excreted in urine mainly in the form of ribo­flavin (60–70% of all urinary flavins) or other metabolites, such as 7‑hydroxy­methyl­ribo­flavin (7‑α-hydroxy­ribo­flavin) and lumi­flavin (Chastain and McCormick, 1987). Very little ribo­flavin can be stored in the body, so urinary excretion reflects dietary intake after tissues become saturated. A strong positive linear correlation of urinary ribo­flavin with ribo­flavin intake above 1.4mg/day has been reported (r² = 0.9667, p < 0.01) (Guo et al., 2016). Urinary ribo­flavin excretion is considered a bio­marker of short-term status.

Experimental balance studies indicate that urinary ribo­flavin excretionary rates increase slowly with increasing intakes until the point of tissue saturation, after which any further increase of ribo­flavin intake leads to a sharp elevation of the excretion rate, as shown in Figure 20b.3.

Based on the weighted results of experimental studies (Brewer et al., 1946; Horwitt et al., 1950; Boisvert et al., 1993b; Guo et al., 2016), the European Food Safety Authority estimated that the change of the urinary excretion rate of ribo­flavin corresponding to the point of tissue saturation could be achieved by a ribo­flavin intake of around 1.3mg/d (EFSA, 2017). Once intakes of 2.5mg/d are reached, excretion becomes approximately equal to the rate of absorption (Horwitt et al., 1950). At such high intakes, a significant proportion of the ribo­flavin intake is not absorbed.

Urinary ribo­flavin was measured in several early population studies in North America: the Nutrition Canada survey, the US Ten State Nutrition Survey, and in NHANES I. In these surveys, casual urine samples were collected rather than the preferable 24h urine samples. As a result, ribo­flavin concen­trations were expressed per gram of creatinine. For children, the rate of ribo­flavin excretion, when expressed per gram of creatinine is greater than for adults. The population-based VERA Study, conducted as part of the German National Consumption Study I, assessed urinary ribo­flavin in 24hour urine samples of 2,006 adults (Heseker et al., 1992)

Occasionally, a ribo­flavin load test is used to assess the degree to which the body is saturated with ribo­flavin. An oral dose of 5mg ribo­flavin is given, and again the timed excretion of ribo­flavin in the urine is measured, usually over a 4h period (Lossy et al., 1951; ICNND, 1963; Brun et al., 1990); the results should be compared with the level in the urine before the test load.

          Factors affecting ribo­flavin excretion

Physical activity and sleep can decrease ribo­flavin excretion (Soares et al., 1993).

Negative nitrogen balance and infection induce a breakdown of tissue protein and, hence, increase urinary excretion of ribo­flavin (Tucker et al., 1960). Consequently, measurement of urinary ribo­flavin is not very useful in circumstances where intakes of protein are low or where chronic infection is common.

Drugs including some anti­bio­tics and psycho­tropic drugs such as pheno­thiazines can increase the excretion of ribo­flavin (Goldsmith, 1975).

Oral contraceptive agents and pregnancy influence ribo­flavin excretion. Concen­trations decrease when oral contraceptive agents are used, and during the third trimester of pregnancy, as tissue retention of the vitamin is increased in response to increased need. However, during the second trimester, excretion increases.

Within-subject variability can be large for 24h urinary ribo­flavin excretion. Coefficients of variation range from 13% to 25%, which are not reduced when results are expressed in terms of creatinine (van Dokkum et al., 1990).

          Interpretive criteria

Interpretive criteria for urinary ribo­flavin excretion for adults and children are shown in Table 20b.4.

Table 20b.4: Interpretive criteria for the urinary excretion of riboflavin. tri: trimester. From (Sauberlich HE ,1999).

Subjects	Less than acceptable (at risk)		Acceptable (low risk)
	Deficient (high risk)	Suboptimal (med. risk)
Children	(µg/g creatinine)
1–3y	< 150	150–499	≥ 500
4–6y	< 100	100–299	≥ 300
7–9y	< 85	85–269	≥ 270
10–15y	< 70	70–199	≥ 200
Adults	< 27	27–79	≥ 80
Preg. 1st tri.	< 27	27–65	66–129
Preg. 2nd tri.	< 23	23–54	55–109
Preg. 3rd tri.	< 21	21–49	50–99
Adults	Other interpretive guidelines
µg/24h	< 40	40–119	≥ 120
µg/6h	< 10	10–29	≥ 30
µg in 4h after    ribofl. load	< 1000	1000–1399	≥ 1400

For adults, when dietary intakes are adequate, ≥ 120µg of ribo­flavin per day should be excreted. After a loading dose of 5mg ribo­flavin, then ≥ 1400µg of ribo­flavin should be excreted in a 4h period under conditions of adequate ribo­flavin intake. In a deficiency state, < l000µg of ribo­flavin will be excreted during the same time period. Interpretive criteria for ribo­flavin excretion in casual urine samples expressed as µg/g creatinine for children, adults, and pregnant women are also given in Table 20b.4.

          Measurement of urinary ribo­flavin

In the past, fluorometric and microbio­logical assays have generally been used to determine ribo­flavin concen­trations in urine. The fluorometric method measures the fluorescence of the flavins directly or converts the flavins to lumiflavin and then determines the fluorescence. The microbio­logical methods use Ochromonas danica or the protozoan Tetrahymena pyriformis.

Measurement of urinary ribo­flavin using HPLC with fluorometry (Gatautis and Naito, 1981) or competitive protein-binding assays (Tillotson and Bashor, 1980) is also possible. The protein-binding assays are rapid, sensitive, and require no treatment of the urine sample prior to analysis. Discrepancies among the analytical methods have been documented. Results for the HPLC method with fluorometry tend to be lower than those with the older fluorometric method alone because, in the HPLC method, ribo­flavin is separated from other flavins (Smith, 1980).

20b.2.3 Ribo­flavin, FMN and FAD in blood

Plasma or serum concentration of ribo­flavin and its derivatives, FMN and FAD have been used in a few studies for assessment of ribo­flavin status. The results of these investigations showed that supple­mentation with low doses of ribo­flavin increased plasma free ribo­flavin (Bessey et al., 1956; Hustad et al., 2002; Guo et al., 2016) and FMN (Bessey et al., 1956; Hustad et al., 2002) whereas no detectable or very modest changes were reported for plasma FAD concen­trations (Bessey et al., 1956; Zempleni et al., 1996; Hustad et al., 2002).

The concentration of ribo­flavin forms in erythrocytes could be considered as marker of long-term ribo­flavin intake. Erythrocyte concentration of ribo­flavin (mainly FMN and FAD) was found to be significantly lower in individuals with clinical signs of ribo­flavin deficiency compared with healthy people (Bamji, 1969). Supple­mentation with a low dose of ribo­flavin for 12wks increased erythrocyte FMN by 87% whereas more modest but still significant response was reported for erythrocyte FAD (Hustad et al., 2002). A good relationship was shown between EGR AC and erythrocyte ribo­flavin (Bates et al., 1999) and erythrocyte FMN and FAD (Hustad et al., 2002) indicating that these measurements have potential and warrant further exploration.

Fluorometric and microbio­logical assays have been used to determine ribo­flavin in plasma / serum and erythrocytes. Hustad et al. (1999) have developed a sensitive and robust method, based on capillary electrophoresis and laser-induced fluorescence detection, for quantifying the low physiological concen­trations of FMN and FAD, as well as ribo­flavin, in human plasma; later the method has been developed further for assessment of ribo­flavin derivatives in erythrocytes (Hustad et al., 2002).

20b.2.4 pyri­dox­amine phosphate oxidase (PPO) activity and activity coefficient

The measurement of erythrocyte PPO enzyme activity and its activity coefficient (PPO AC; assessed by the ratio of PPO activity before and after in vitro activation with the prosthetic group FMN) have been suggested as alternatives to EGR AC for assessment of ribo­flavin status. This option could be very useful for diagnosis of ribo­flavin deficiency in populations with a high prevalence of glucose-6-phosphate dehy­drogenase deficiency since the EGR AC assay produces erroneous results in individuals with this genetic defect (Thurnham, 1972; Prentice et al., 1981).

A fluorometric assay for measurement of PPO activity in red blood cell lysates has been developed (Bates and Powers, 1985) and subsequently optimized and adapted for assessment of PPO AC (Mushtaq et al., 2009). The results generated by this methodology demonstrated that both PPO and PPO AC are responsive to ribo­flavin supple­mentation and PPO showed a strong dose-response effect. In addition, in people free from glucose-6-phosphate dehy­drogenase deficiency, PPO and PPO AC strongly correlated with the well-established bio­marker of ribo­flavin status, EGR AC (Mushtaq et al., 2009). Therefore, both PPO and PPOAC are promising bio­markers, however, criteria to assess ribo­flavin adequacy based on these bio­markers have not been developed yet.

Acknowledgements

KP would like to gratefully acknowledge the original contribution of Rosalind Gibson on which some parts of this chapter are based and to express gratitude to Mike Jory and Ian Gibson for the HTML design and work in the translation to this HTML version. The assistance of Nutritional International for work on this chapter is also gratefully acknowledged.