| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Department of Clinical Biochemistry, Macewen Building, Royal Infirmary, Glasgow G4 0SF, United Kingdom.
a Author for correspondence. Fax 44-0141-553-1703.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We used reversed-phase HPLC with postcolumn derivatization with alkaline potassium ferricyanide and fluorescence detection. Samples were deproteinized and injected directly onto a C18 column. TDP concentrations in erythrocytes were compared with those in whole blood. Reference intervals for erythrocyte TDP (n = 147; 79 males and 68 females; mean age, 54 years) and whole blood TDP (n = 124; 68 males and 56 females; mean age, 54 years) were determined in an apparently healthy population. We compared erythrocyte TDP with results of the erythrocyte transketolase activation test in 63 patients who were considered at risk of thiamin deficiency.
Results: The method was linear to at least 200 µg/L. The between-run CV was <8%. The lower limit of quantification for both whole blood and packed erythrocytes was 300 pg on column with a detection limit of 130 pg on column. Recovery of TDP from blood samples was >90%. TDP in erythrocytes correlated strongly with that in whole blood (r = 0.97). Reference intervals for erythrocyte and whole blood TDP were 280–590 ng/g hemoglobin and 275–675 ng/g hemoglobin, respectively. Of the 63 patients suspected of thiamin deficiency, 46 were normal by both TDP and activation tests, 13 were deficient by both tests, 1 was deficient by the activation test but had normal erythrocyte TDP concentrations, and 4 were normal by the activation test but had low TDP.
Conclusions: The HPLC method is precise and yields results similar to the erythrocyte activation assay.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The most widely used method to detect thiamin deficiency is the indirect measurement of TDP in erythrocytes with either the transketolase activation test or the transketolase activity assay (5)(6), but these tests are functional rather than a direct measurement of thiamin status and therefore may be influenced by factors other than thiamin deficiency. These include loss of reactivatable apoenzyme during chronic deficiency in vivo (7)(8), altered binding of apoenzyme and coenzyme because of the presence of isoenzymes of transketolase (9), and reduced synthesis of apoenzyme in patients with diabetes (10) and liver disease (11). Other factors that have reduced the usefulness of the transketolase assays have been the relatively poor interassay precision of the assay, difficulty with standardization, and sample storage instability because of rapid inactivation of the transketolase enzyme (7). These factors contribute to the lack of agreement over the upper limit of the reference range for the activation assay, with values ranging from 15.5% to 40% (12)(13).
The TDP concentration in erythrocytes has been shown to be a good indicator of body stores because it depletes at a rate similar to those of other major organs (14). Studies by Warnock et al. (15) and Baines and Davies (16) have suggested that erythrocyte TDP concentrations determined by the apoenzyme recombination technique or by HPLC are a more sensitive index of thiamin status than the measurement of erythrocyte transketolase activity. The use of HPLC for the direct determination of TDP has clear advantages in terms of sensitivity, specificity, precision, and robustness (15)(16)(17).
We describe a simple, robust reversed-phase HPLC method with postcolumn derivatization that is suitable for the routine measurement of TDP in erythrocytes or in whole blood. The method is fast and has adequate sensitivity and precision for use in the assessment of thiamin status in clinical practice and for research purposes. Measurement of TDP in erythrocytes is compared to that in whole blood. Direct measurement of TDP in erythrocytes is compared with the indirect measurement of TDP by the transketolase activation assay as an indicator of thiamin status in patients at risk of thiamin deficiency.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
chromatographic conditions
The HPLC system consisted of a Waters solvent delivery system and
a Waters 470 fluorometer (Waters) The isocratic mobile phase for
chromatography was 20 mmol/L sodium dihydrogen phosphate buffer (pH
4.6) containing 10 g/L acetonitrile. The mobile phase was filtered
through a 0.45 µm nylon filter and pumped through a 5 µm
reversed-phase column (Apex ODS; 4.6 x 250 mm, protected with a
10-mm guard column) at a flow rate of 1.5 mL/min.
The mobile phase was optimized with respect to stability of signal, column life, analysis time, reproducibility of results, and separation by varying its pH and the concentration of acetonitrile and phosphate buffer.
detection
For postcolumn derivatization of TDP, the column effluent was
reacted with 2 mol/L sodium hydroxide containing 600 mg/L potassium
ferricyanide at a flow rate of 0.5 mL/min. The effluents were excited
at 375 nm; the emission was measured at 430 nm and quantified by
measuring the peak height.
preparation of blood samples
Venous blood was collected into heparin-containing tubes.
The packed erythrocytes were washed three times with 9 g/L NaCl, with
care being taken to remove the buffy coat because leukocytes are a rich
source of TDP. To hemolyze the cells, we added 1 mL of distilled water
to 1 mL of packed cells and vortex-mixed the mixture for 1 min. The
hemolysates were stored at -70 °C until analysis. For the
measurement of TDP in whole blood, 1 mL of heparinized whole blood was
stored at -70 °C until analysis.
Blood samples for population reference values were obtained from laboratory staff and from people attending a cardiovascular risk clinic at Glasgow Royal Infirmary. None of the subjects were taking any vitamin supplements or had any significant medical history. Although their dietary data were not recorded, all subjects indicated having reasonably typical diets. Erythrocyte TDP concentrations were measured in 147 subjects, 79 males and 68 females, with an average age of 54 years (range, 22–71 years). The reference range for whole blood TDP was established on 124 subjects, 68 males and 56 females.
preparation of calibrators
A stock solution of TDP was prepared by dissolving 100 mg of TDP
in 100 mL of 0.1 mol/L HCl. This solution was stable for 4 months when
stored at -70 °C. The stock solution was diluted 1:1000 in
distilled water to prepare the intermediate stock solution.
Calibrators were prepared by adding the intermediate stock solution to
a solution of 80 g/L bovine serum albumin in distilled water to provide
a range of 20–80 µg/L TDP. Although the working solution of TDP in
0.1 mol/L HCl was stable for <2 weeks when stored at -70 °C, its
stability was improved when prepared in 80 g/L bovine serum albumin
solution. Calibrators stored were stable for 4 weeks when stored at
-70 °C.
quality control
Internal quality control (QC).
Hemolysates were prepared as
described above. QC1 was the hemolysate, and QC2 was prepared by
addition of the intermediate TDP solution to the hemolysate. The QC
hemolysates were aliquoted and were stable for at least 12 months when
stored at -70 °C.
External quality assessment.
Currently there is no external
quality assessment scheme available in the United Kingdom for
measurement of TDP by HPLC.
sample preparation
Samples and QC hemolysates were thawed, vortex-mixed, and
centrifuged at 1000g for 10 min to remove cell debris.
Distilled water (500 µL) was added to 500 µL of whole blood, red
cell, or QC hemolysate in a small centrifuge tube and vortex-mixed for
1 min; 500 µL of this diluted hemolysate was transferred to another
tube, and 75 µL of 500 g/L TCA was added to precipitate the protein.
The precipitated sample was thoroughly vortex-mixed and centrifuged for
10 min. A 200-µL aliquot of the supernatant was transferred to
conical autosampler vials, and 50 µL was injected onto the HPLC
column via an autosampler.
The remainder of the diluted hemolysate was used for hemoglobin (Hb) estimation using Drabkins reagent (Sigma Diagnostics), which contains potassium ferricyanide and potassium cyanide. Hb is oxidized and converted to stable cyanmethemoglobin by ferricyanide and cyanide, respectively. The absorbance of cyanmethemoglobin was measured at 540 nm using the Cobas Mira analyzer (Roche Products).
calculation of results
The calibration curve was prepared on the day of analysis by the
addition of 75 µL of 500 g/L TCA to 500 µL of calibrators
(containing 20, 40, and 80 µg/L TDP), which were treated in the same
way as the sample hemolysates. Six calibrators were run with each batch
of 30 samples: two at the start, two in middle, and two at the end of
the run.
stability studies
To check the stability of TDP in whole blood at room temperature,
five different whole blood samples were treated freshly and after 24,
48, and 72 h. Each blood sample was analyzed three times to obtain
mean TDP values at the times specified above. The long-term storage
stability of hemolysates and whole blood was checked by aliquoting a
pooled hemolysate and a whole blood sample. The aliquots were stored at
-20 °C and -70 °C and analyzed over a period of 7 months.
transketolase activation assay
The transketolase activation effect was measured by adapting the
method of Bayoumi and Rosalki (18) for use on the Cobas Bio
centrifugal analyzer (Roche Products) as described by Mount et al.
(19). The activation result was calculated by expressing the
two activities of transketolase enzyme (with and without addition of
TDP cofactor) as a percentage. The upper reference limit for
the activation assay was obtained by determining the activation effect
on blood samples from apparently healthy subjects (n = 79). The
upper cutoff established in our laboratory was 25% (mean + 2 SD).
The within-batch imprecision (CV) of the assay, which was obtained by analyzing 17 aliquots of the same hemolysate, was 5.9%. The between-batch CV was 21%, as assessed by analyzing a hemolysate aliquoted and stored at -70 °C on 10 occasions over 2 weeks. Others have reported similar CVs, which are a reflection of sample instability during storage of hemolysates and show deterioration of transketolase activity over a 2-week period when stored at -70 °C (7)(19).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, A–C. The TDP peak was well resolved and identified by
comparing peak retention time with that of an albumin solution with
added TDP. The TDP peak eluted after 4 min, and because there were no
later eluting peaks, further injections were made at intervals of 6
min. The chromatographic profile of a blank injection prepared by
omitting the thiochrome reaction suggested a lack of interference in
sample extracts (Fig. 1D
). Under the chromatographic conditions
described, there was no interference from either TMP or thiamin, which
together with thiamin triphosphate are present in very low
concentrations in whole blood (<5% of total thiamin). TMP has a
retention time of 6.2 min, with thiamin eluting after 9.0 min. With
this chromatographic system, we detected neither thiamin nor TMP in
either whole blood or red cells.
|
The concentration of the phosphate buffer in the mobile phase affected column stability and performance. At concentrations >100 mmol/L, the column was unstable, and at <10 mmol/L, the TDP peak was broader and retention times were longer. A phosphate buffer concentration of 20–50 mmol/L avoided both problems. The intensity and reproducibility of the fluorescence signal was found to depend on acetonitrile as an organic modifier in the mobile phase. A final concentration of 10 g/L acetonitrile in the mobile phase yielded a TDP peak that was sharp and well resolved with reproducible retention time and signal response. Varying the pH of the mobile phase between 3.0 and 6.0 did not affect the retention time, resolution, or intensity of the TDP signal.
For postcolumn derivatization of TDP, 0.6 g/L potassium ferricyanide in 2 mol/L sodium hydroxide solution delivered at a flow rate of 0.5 mL/min produced an optimal signal response with minimum peak broadening.
analytical validation
Linearity and detection limit.
The calibration curve was
linear up to least 200 µg/L. The minimum detectable concentration,
defined as the lowest concentration of the analyte in the calibrator
that could be detected from zero with 97% confidence (n = 9) was
3 µg/L TDP (130 pg on column). The limit of quantification,
defined as 10 times the signal-to-noise ratio, was 7 µg/L TDP (300 pg
on column) for both whole blood and packed red cells.
Recovery studies.
When 50 and 100 µg/L TDP was added to a
pool of washed red cells, the recoveries were 95% ± 6% (mean ±
SD; n = 5) and 97% ± 4.8% (n = 5), respectively. The
corresponding recoveries for a whole blood sample were 97% ± 5.0%
(n = 5) and 93% ± 6.7 (n = 5), respectively.
Precision studies.
Precision data are shown in Table 1
. The within-batch imprecision for the measurement of TDP in red
cells was calculated by analyzing QC1 and QC2 10 times on the same day.
The between-batch imprecision data were obtained from QC1 and QC2
material analyzed 23 times over a period of 5 months.
|
The within-batch precision data for the measurement of TDP in whole blood were obtained from a pooled whole blood sample analyzed 10 times on the same day. Between-batch imprecision was obtained by analyzing the pooled whole blood sample aliquoted and stored at -70 °C over a period of 3 months (n = 14).
Recalibration of assay.
Within the same HPLC batch, the assay
was recalibrated after every 15 samples to correct for any drift and to
monitor the stability of the chromatographic system. No significant
drift was observed in calibrators at the start and end of a batch
containing 30 samples.
stability studies
Hemolysates and whole blood stored at -70 °C were stable for
at least 7 months. The mean value of a pooled hemolysate at the start
of a 7-month storage period was 372 ng TDP/g Hb (range, 355–391 ng/g
Hb; n = 5), and at the end, it was 384 ng TDP/g Hb (range,
368–405 ng/g Hb; n = 5). The corresponding mean values for whole
blood were 404 ng TDP/g Hb at the start (range, 377–421 ng/g Hb;
n = 5) and 387 ng TDP/g Hb after 7 months (range, 370–407 ng/g
Hb; n = 5). Although hemolysates were stable at -20 °C for at
least 7 months, the mean TDP value in whole blood stored at -20 °C
was significantly different (P <0.01) from that at the
start (mean, 331 ng TDP/g Hb; range, 258–377 ng/g Hb; n = 5). We
therefore stored all blood samples at -70 °C. At room temperature,
whole blood was stable for TDP for up to 48 h. After 72 h, a
variability of -10.3% to 13.7% was observed. Differences between
mean TDP values in whole blood after 72 h and when freshly treated
were significantly different for all five blood samples analyzed
(P <0.05 to <0.01).
TDP in deproteinized samples was stable in the autosampler for up to 12 h at room temperature and for up to 24 h when the autosampler was refrigerated at 10 °C.
population reference values
Whole blood and red cell TDP values showed a gaussian
distribution. The Kolmogorov–Smirnov test revealed no significant
departures from a gaussian distribution (P >0.15). The mean
concentration of TDP in red cells was 433 ng/g Hb (n = 147) with a
95% reference interval of 280–590 ng/g Hb (mean ± 2
SD). There was no significant difference in TDP concentrations
between males and females. The corresponding reference interval for
whole blood TDP was 275–675 ng/g Hb with a mean of 460 ng/g Hb (n
= 124).
comparison studies
Transketolase activation assay vs HPLC red cell TDP assay.
Sixty-three acute and chronic medical and surgical patients who were
considered to be at risk of thiamin deficiency were assessed for
thiamin status by both HPLC and transketolase activation assays.
Clinical assessments by the two methods agreed in 58 of the 63 patients
(Fig. 2
). Fourteen patients were considered to be thiamin deficient by
the transketolase activation test, and 13 of these patients also had
low red cell TDP by the HPLC method. However, four patients with low
TDP concentrations had transketolase activation values within the
reference range. Activation values in these patients were
21–23%, which is close to the upper cutoff for biochemical deficiency
established by the transketolase activation assay (>25%).
|
Chronic renal failure patients are at risk of thiamin
deficiency as a consequence of dietary restrictions and loss of thiamin
during dialysis. The thiamin status of six chronic renal failure
patients was assessed by both assays in an attempt to establish whether
thiamin supplements were adequate. All patients were undergoing
intermittent dialysis and thiamin supplementation. Both the activation
and HPLC assays showed that no patients were thiamin deficient either
pre or post dialysis. However, the HPLC assay provided the additional
information of thiamin accumulation in the red cells of these patients,
which is cleared during dialysis (Table 2
).
|
Red cell TDP assay vs whole blood TDP assay.
TDP
concentrations in red cells and in whole blood were measured by HPLC in
76 patients who were receiving nutritional support. Measurement of TDP
in red cells correlated strongly with that in whole blood
(r = 0.97). The regression equation was: whole blood
TDP (ng/g Hb) = 11.2 + 0.97 x red cell TDP (ng/g Hb). A
difference plot (20) also showed that there was good
agreement between red cell and whole blood TDP concentrations (Fig. 3
). The mean difference ±2 SD was 27 and -25 ng/g Hb,
which represents an 8% difference between the two assays at the lower
end (300 ng TDP/g Hb) and 3% at the upper end (800 ng TDP/g Hb). The
95% confidence interval for the mean difference was 4.4 to -1.8 ng
TDP/g Hb, indicating that there was no detectable bias.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We found that erythrocyte TDP concentrations measured by HPLC compared
well with the erythrocyte transketolase activation assay when assessing
patients with suspected thiamin deficiency. Other workers who used HPLC
with a precolumn derivatization procedure have reported similar results
(15)(16). However, four of our patients who had
activation values just below the upper cutoff (<25%) were clearly
deficient by HPLC (Fig. 2
), which may indicate that measurement of TDP
in erythrocytes is a more sensitive indicator of thiamin status than
the transketolase activation assay. This has also been suggested by
Warnock et al. (15) and Baines and Davies (16).
Unlike the activation test, the HPLC assay can also detect tissue
thiamin accumulation, as demonstrated in patients on hemodialysis for
chronic renal failure.
Blood and tissue concentrations of thiamin TDP are not subject to any known physiological regulation but are dependent on dietary intake. Diets are very variable globally, being influenced by local agriculture and economy and increasingly, in affluent societies, by fashion. Thus, the concepts of intra- and interindividual biological variation are not applicable. Recommendations for optimal intake of vitamins to maintain health such as the Recommended Dietary Allowances in the US (21) and the Reference Nutrient Intake values in the United Kingdom (22) have been published, but although similar, they have some significant differences and are not universally accepted. It is therefore essential that a local reference range be established. Our population reference range for TDP in red cells is comparable with those reported previously (17)(23)(24). However, we related TDP concentrations in erythrocytes to Hb rather than to volume of packed red cells because accurate pipetting of packed red cells was difficult because of their viscosity, which adversely affected the precision of the HPLC assay. Relating the erythrocyte TDP concentration to Hb in the sample improved the reproducibility of the assay.
Measurement of TDP concentrations in red cells correlated strongly with that in whole blood. TDP concentrations in red cells and in whole blood showed good agreement when corrected for Hb. This is not surprising because TDP is present predominately in red cells (>90%). Because whole blood is convenient to use and preparation of washed red cells is time-consuming and has safety implications, we recommend the measurement of TDP in whole blood rather than washed red cells for the laboratory assessment of thiamin status. However, whole blood analysis should be carried out only if accompanied by Hb estimation to correct for any cell volume variability.
HPLC methods for the measurement of total thiamin or TDP in blood can be divided into pre- and postcolumn derivatization techniques with or without hydrolysis of TDP (3)(6)(16)(17)(24)(25)(26)(27)(28)(29)(30)(31). Measurement of total thiamin in whole blood requires time-consuming enzymatic hydrolysis of TDP to free thiamin before HPLC analysis (3)(6)(27). The sensitivity and precision of the postcolumn derivatization method described here for the determination of TDP in erythrocytes or whole blood compare favorably with precolumn methods that have been described previously (16)(17)(23)(24). Although both techniques are simple to perform, the postcolumn procedure is more robust and less labor-intensive and time-consuming (analysis time <15 min) than precolumn methods, which usually require removal of TCA from the acid extracts with ether (15)(23) or its neutralization with sodium hydroxide (24)(28) before derivatization with alkaline potassium ferricyanide. In addition, the precolumn derivatization procedure uses relatively harsh chromatographic conditions because of the need to use an alkaline mobile phase (pH 7.5–8.0) for the thiochrome derivatives to fluoresce. In our experience, use of such mobile phases leads to fairly rapid loss of column performance because of the dissolution of silica at high pH. Other disadvantages include poor yield of reaction (<67%) and instability of thiochrome derivatives at pH values below 8.0 (32). These problems are largely eliminated with the use of postcolumn derivatization, as described here and by Kimura and Itokawa (25), where the sample is deproteinized and injected directly onto a reversed-phase column with subsequent derivatization of the eluting substances. However, this procedure requires use of additional equipment (a postcolumn pump and a reaction coil) and can lead to peak broadening.
The HPLC method described here is suitable for the measurement of TDP in both whole blood and red cells. The method is simple and has adequate sensitivity, selectivity, and precision for use in the routine laboratory assessment of thiamin status in patients suspected of thiamin deficiency. The reference range for erythrocyte TDP is dependent on dietary thiamin intake because there are no known physiological mechanisms that regulate its concentration in blood or tissues. Thus, the erythrocyte TDP concentration in any individual will span the reference range according to his or her dietary intake. The between-batch CV of the proposed method (<8%) is much less than one-half the biological variation and therefore fulfils the current criteria (33) for suitability in clinical use.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. R. Matrana, S. Vasireddy, and W. E. Davis The Skinny on a Growing Problem: Dry Beriberi after Bariatric Surgery Ann Intern Med, December 2, 2008; 149(11): 842 - 844. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Lu and E. L. Frank Rapid HPLC Measurement of Thiamine and Its Phosphate Esters in Whole Blood Clin. Chem., May 1, 2008; 54(5): 901 - 906. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Hanninen, P. B. Darling, M. J. Sole, A. Barr, and M. E. Keith The Prevalence of Thiamin Deficiency in Hospitalized Patients With Congestive Heart Failure J. Am. Coll. Cardiol., January 17, 2006; 47(2): 354 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. K. Talwar, M. K. Azharuddin, C. Williamson, Y. P. Teoh, D. C. McMillan, and D. St. J. O'Reilly Biological Variation of Vitamins in Blood of Healthy Individuals Clin. Chem., November 1, 2005; 51(11): 2145 - 2150. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. McLaren Howard, D. Talwar, and D. St. JO'Reilly Assessment of Vitamin B1 Status Clin. Chem., November 1, 2000; 46(11): 1867 - 1868. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
