QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Erhardt, J. G. Articles by Biesalski, H. K. Search for Related Content PubMed PubMed Citation Articles by Erhardt, J. G. Articles by Biesalski, H. K.

---

Research Communication

Rapid and Simple Measurement of Retinol in Human Dried Whole Blood Spots^{1 ,2}

Juergen G. Erhardt^*3, Neal E. Craft, Felix Heinrich^* and Hans K. Biesalski^*

^* University of Hohenheim, Institute of Biological Chemistry and Nutrition, 70599 Stuttgart, Germany and Craft Technologies, Inc., Wilson, NC 27893

³To whom correspondence should be addressed. E-mail: erhardtj{at}uni-hohenheim.de

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 
We describe an improved method for the measurement of retinol in dried blood spots (DBS) on filter paper. Retinol in human DBS on filter paper was analyzed by normal phase HPLC after a simple extraction method. Retinol associated with its binding protein was eluted from the paper into aqueous solution facilitated by ultrasonic agitation. Retinol associated with retinol binding protein was denatured with acetonitrile, and then retinol was isolated in a single hexane extract and analyzed directly by HPLC. When analyzing DBS, the individual plasma volume of the spots was calculated by measuring the sodium content or by weighing the blood spots. The described method yielded low intra- and interassay variability (<6%), with sufficient sensitivity (detection limit, 0.1 µmol/L) and good recovery (97% spike). Compared with matching plasma samples, DBS retinol consistently decreased 18–23% during the 1st wk of storage. After 1 wk, retinol remained stable in the blood spots at 23°C for >3 mo. In conclusion, the analysis of retinol in DBS by HPLC is comparable to retinol analysis in serum. The variability of the method was reduced by using sodium concentration to estimate sample volume. Collection of DBS for retinol analysis is appropriate under field conditions, where it is difficult to centrifuge or freeze blood samples.

KEY WORDS: • retinol • vitamin A status • dried blood spot • HPLC • filter paper • nutritional assessment

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 
Worldwide, vitamin A deficiency (VAD)⁴ is one of the most important dietary deficiencies. Each year it is responsible for >350,000 cases of blindness in children in addition to significant increases in childhood and maternal morbidity and mortality among the poor of the nonindustrialized world. In all, >250 million children are subclinically affected while 5 million exhibit clinical symptoms (1 ). In recent years studies have shown that vitamin A supplementation can reduce the mortality of children by up to 50% (2 ). With these facts in mind, simplifying the assessment of vitamin A status is an important analytical challenge in nutritional science.

Several methods are available to measure VAD. The determination of retinol in blood is one of the most frequently used methods and was also recommended by the World Health Organization in 1994 (3 ). During normal vitamin A intake, plasma retinol level is closely regulated; however, during manifest deficiency the level of retinol decreases and is a reliable marker of the vitamin A status if individuals with infection are excluded. The disadvantages of published HPLC methods include: the requirement for relatively large amounts of venous blood (>0.2 mL) and the storage of samples below 0°C under field conditions.

To overcome these limitations blood spots can be collected from a finger-prick and stored at up to 25°C before analysis. For decades this method of sample collection has been well-established for the measurement of several substances in blood, e.g., phenylalanine to screen for phenylketonuria (4 ).

Previously, we have demonstrated that it was possible to reliably measure retinol in dried blood spots (DBS) (5 ,6 ). Unfortunately, the sample preparation was rather laborious and the recovery of retinol was less than ideal (55–65%) making the calculation of retinol in dried whole blood spots difficult. To compensate for the low recovery, a small number of representative, matching, plasma samples have been collected to calculate an "adjustment factor." This factor accounted for the unknown volume in a 6.35-mm diameter circle, changes during storage and recovery during DBS extraction. The need for a subset of matching plasma samples and the time-consuming sample preparation greatly hampered the advantages of the blood spot technique.

The objectives of this study were to further improve the measurement of retinol in DBS including storage at ambient temperature, simplified extraction and improved HPLC methodology; to evaluate methods to account for the plasma content of DBS, including sodium content and DBS weight and, thus, improve the calculation of retinol concentration; and to test the new procedure in DBS from India, Nicaragua and Indonesia.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 
Chemicals.

The following chemicals were used as purchased: acetonitrile, hexane, isopropyl alcohol (HPLC-Grade; Merck, Darmstadt, Germany), tocol (Hoffman-La Roche, Basel, Switzerland), retinol, 2,6-ditert-butyl-p-cresol (BHT), ascorbic acid, Triton X 100, (p.a. grade; Sigma, Aldrich, Steinheim, Germany), lithium chloride (Eppendorf, Hamburg, Germany), National Institute of Standards and Technology (NIST) standard reference material 968b (National Institutes of Standards and Technology, Gaithersburg, MD).

Sample collection.

Blood letting was carried out using Microtainer Safety Flow Lancets (Becton Dickinson, Franklin Lakes, NJ). The skin was warmed before puncture to increase the amount of blood obtained.

To prepare the blood spots, 30 µL of blood were deposited directly on a filter paper (Schleicher and Schuell 903 specimen collection paper; Dassel, Germany) and subsequently dried for 3 h in the dark. The spots were stored with a desiccant in the dark at room temperature (23°C) in Ziploc bags.

Validation procedures.

For validation purposes, retinol from three subjects was measured in DBS that had been stored at least 1 wk before analysis. The intra- (n = 6) and interassay variances (n = 5) were determined during a 5-wk period. Linearity and spike recovery was assessed by adding 5 progressively increasing concentrations of retinol (0.5–2.5 µmol/L) to whole blood samples before extraction.

Effect of handling and storage at ambient temperature.

Whole blood was collected from three subjects and either frozen at -80°C or dried and stored in the dark at room temperature for 3 mo. Retinol was measured at 0, 1, 2, 3, 6, 12, 30, 60 and 90 d in DBS and compared with retinol measured in whole blood samples thawed and measured at the same time-points. To test different handling conditions, additional samples were divided in four groups. The first set of blood spot collection cards was stored without desiccant in a Ziploc bag after drying for 1 h. The second set was dried the same amount of time but stored with desiccant. The third set was dried 12 h and then stored without desiccant, and the fourth set was dried the same but stored with desiccant. To test the applicability of the new procedure DBS from India, Nicaragua and Indonesia, matching plasma samples were analyzed for correlation and recovery. The individuals provided informed consent and the study was performed in accord with the 1983 revision of the Helsinki Declaration.

Extraction of retinol from DBS.

A 6.35-mm diameter circle (12 µL blood) from the DBS or the whole spot was extracted in 500 µL distilled water containing 10 g/L ascorbic acid, as an antioxidant, for 15 min followed by 5 min of sonication. Optionally, 0.5 g/L of Triton X 100 was added as a detergent. Subsequently, 400 µL acetonitrile (containing 5 g/L BHT and 4 µmol/L tocol as internal standard) were added and briefly mixed. Next 400 µL n-hexane (containing 5g/L BHT) were added and the retinol was extracted for 2 min by vigorous shaking or vortex-mixing. The mixture was centrifuged for 1 min at 8000 x g and 200 µL of the supernatant were injected.

HPLC analysis of DBS.

The HPLC consisted of a standard HPLC pump, a programmable UV detector and an autosampler (Thermo Separation Products, San Jose, CA). The separation was achieved using a BDS Hypersil CN 150-mm, 5-µm column in combination with a Javelin NH₂ guard column (Keystone Scientific, Bellefonte, PA). An isocratic mobile phase consisting of hexane/isopropanol (98.5:1.5) was used in recirculation mode for 250 samples. The flow rate was maintained at 1.5 mL/min. From 0 to 2.5 min the wavelength was set at 325 nm and switched to 300 nm from 2.5 to 4 min to get the maximum absorbance of retinol and tocol. Alternatively, a constant wavelength of 300 nm can be used. A diode array detector was used to check the purity of the peaks. The NIST Standard Reference material SRM968b and/or neat solutions of retinol were used for calibration.

	Calculation of plasma volume in DBS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 
Sodium determination.

After the retinol had been extracted, the remaining organic components in the extracted samples were removed by drying overnight in an oven at 60°C. The remaining unevaporated aqueous residue was diluted to 2 mL with distilled water and 20 µL of an aqueous solution of lithium chloride (10 g/L) was added. After centrifugation (8000 x g for 1 min) the sodium content in the supernatant was measured by flame photometry according to the description of the manufacturer (Eppendorf, Hamburg, Germany). To calculate the amount of plasma in the DBS, a value of 140 mmol/L sodium in plasma was assumed (7 ).

Determination based on spot weight.

DBS were spotted with graded volumes of blood (20, 40, 60 and 80 µL) and dried as described above. The hematocrit was measured in the blood so that the volume of plasma in each DBS could be determined. The entire spot was cut from the collection card and weighed before extraction. The relationship between spot weight and plasma volume was determined.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 
Chromatograms.

Figure 1 shows a chromatogram of a DBS. Both peaks (retinol and tocol) were baseline separated with an analysis time of 4 min.

View larger version (14K): [in this window] [in a new window]   Figure 1. Chromatogram of dried blood spots (DBS) retinol (0.6 µmol/L) from a sample collected in India containing 50 µL of whole blood. The mobile phase was hexane/isopropanol (98.5:1.5) with a flow rate of 1.5 mL/min (detection at 325 nm from 0 to 2.5 min and 300 nm from 2.5 to 4 min).

The detection limit in HPLC is defined to be the amount injected that generates a peak height threefold that of baseline noise. For retinol, 0.1 µmol/L was the detection limit when 20 µL whole blood was applied to the filter paper. The assay was linear (r² = 0.9995) to at least 2.5 µmol/L as determined by the addition of neat solutions of retinol to DBS. The mean recovery of spiked solutions was 97 ± 4%.

In unspiked blood spots from three subjects, the intra- (n = 6) and interassay (n = 5) variances (CV) were <6% for retinol.

Stability and recovery from the DBS.

During the 1st wk of DBS storage retinol concentration in each sample decreased 18–23% (Fig. 2 ). After the 1st wk retinol remained stable at room temperature (23°C) for at least 3 mo.

View larger version (16K): [in this window] [in a new window]   Figure 2. Effect of storage time at room temperature (23°C) on dried whole blood spot retinol from three subjects.

The samples dried for 1 h and stored with desiccant and samples dried 12 h stored with or without desiccant (n = 4 for each) has similar recoveries of retinol which were 77–82% of the level in the same amount of whole blood stored at -80°C. Only in the samples stored without desiccant after 1 h of drying in Ziploc plastic bags was the recovery reduced to 68%.

Application of the method to samples from India, Nicaragua and Indonesia.

A high correlation (r² > 0.95) between retinol in DBS and matching plasma samples was observed for samples from developing countries (India, n = 12; Nicaragua, n = 10; Indonesia, n = 24;Fig. 3 ). After correcting the measured DBS retinol for the serum volume in the DBS based on sodium, the recovery in these samples was in the range of 75–84%.

View larger version (11K): [in this window] [in a new window]   Figure 3. Correlation of plasma retinol concentrations with dried blood spot (DBS) retinol values in 46 samples from three different countries (Nicaragua, India and Indonesia). DBS retinol was measured after >1 wk of storage at room temperature.

The intra- (n = 6) and interassay variance (n = 4) of the measurement of sodium was always <5%. Differences between whole blood sodium measured directly and sodium extracted from DBS samples were <6% in all samples (n = 6). The measurement of sodium in the DBS extracts was linear (r = 0.99) without further dilution for 5–40 µL volumes of whole blood applied to the filter paper.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 
VAD is often prevalent in very remote areas where the poor availability of electricity and facilities make it difficult to centrifuge or refrigerate blood samples. In addition, it is often impossible to obtain venous blood samples from specific populations. Under such circumstances, methods must be adopted where small volumes of blood, for example, from a finger-stick, can be used. A reasonable approach is to collect and dry small blood samples on filter paper before extraction and analysis of blood components. Published methods to measure retinol on filter paper have several limitations, including necessity of larger plasma volume (8 ), necessity of expensive equipment (9 ) or the need for a subsample of matching plasma samples to account for recovery and correctly calculate the plasma retinol values (6 ).

Using the modified method, we achieved very reproducible results that were comparable to our previous extraction procedure and reversed-phase HPLC separation (6 ). The addition of Triton X 100 to the extraction solvent as a detergent slightly reduced the CV by 1%, but it tended to generate stable emulsions. High-speed centrifugation is useful for breaking this emulsion. The use of Triton is optional and should be avoided if high-speed centrifugation is not available. As for any hexane extraction of aqueous solvents, efficient mixing is an important step and should be tested thoroughly. In general, high-speed vortexing or vigorous vertical shaking of the tubes for 2 min is sufficient to provide full recovery. Several benefits resulted from the change to a normal-phase HPLC method: 1) the separation of retinol from other interfering peaks was much better, 2) the intermittent column flushing, as previously reported, was not necessary and 3) the laborious evaporation step could be omitted. These advantages greatly increase sample throughput in large-scale epidemiological surveys.

Tocol was chosen as an internal standard because it is well separated from retinol under the normal phase conditions. Although retinyl acetate may have been a logical choice due to its structural similarities and commercial availability, it does not separate well from early eluting components under these conditions. Tocol has been used successfully as an internal standard for retinol analysis by several others and is currently commercially available (Matreya Pleasant, Gap, PA).

Another major observation is that retinol tends to form a gradient across the DBS. In many samples the red blood cells were denser in the center while serum components were higher at the periphery. For this reason we looked for a means of normalizing each sample. Because sodium is tightly regulated (140.0 ± 2.4 mmol/L) (7 ) and like retinol, is primarily present in serum, it seemed to be a reasonable choice. We have observed that sodium distributes similarly to retinol in DBS. Furthermore, the sodium concentration is assumed to be constant except under rare conditions, which would also tend to influence the retinol content in a similar manner, e.g., dehydration. The determination of sodium concentration using flame photometry is easy, reliable and fast. Flame photometers are moderately expensive instruments and may not be available at all sites with HPLC equipment. Currently we are investigating the use of sodium selective electrodes as an alternative to the flame photometer. Otherwise, the volume of plasma in DBS can be estimated by three other procedures. First, a known amount of blood could be applied to the filter paper at the time of collection. This approach makes the sample collection less convenient. A second is to punch a hole from the DBS with a defined volume of blood. This is the method used in former publications, but the plasma volume is variable depending on the saturation of the paper and the distribution of plasma in the DBS. The third is to cut out and weigh the whole spot. The latter is accurate if a calibration curve has been established by spotting different volumes of blood. Although there are alternatives, the determination of the sodium concentration to calculate the plasma content of the DBS is more exact because the sodium concentration of plasma (95% confidence interval: 135–145 mmol/L) is much more constant than the hematocrit (95% confidence interval: 0.39–0.55) (7 ). Under field conditions it would be more convenient to apply an ample but unknown volume of whole blood onto the filter paper and to calculate the plasma content afterwards. Cutting and weighing each spot is much more tedious than using a hole punch. A major advantage of using the entire spot is that higher sensitivity is obtained and uneven distribution of retinol throughout the whole spot is avoided. As mentioned, we have observed higher concentrations of retinol at the periphery of the spot than in the interior. This explains why previously we reported recovery of 55% with center punches compared with 80% at this time using the whole spot. This gradient effect has also been observed by Cook et al. (10 ) for transferrin receptor in DBS.

We observed no difference between the retinol content of whole blood samples stored at -80°C and DBS dried at room temperature for 3 h and analyzed immediately. As reported previously, a decrease in DBS retinol concentration of 20% (18–23%) was observed during wk 1. Fortunately, after this initial decrease, the retinol content in the samples investigated here remained stable when they were stored for 6 mo at ambient temperature (23°C). This phenomenon of degradation or reduced extraction efficiency during wk 1 and the stability in the following weeks is difficult to explain. With our current procedure we observed consistent recovery of 80% in all samples. When DBS on filter paper were stored for at least 1 wk before analysis, the values could be adjusted by 20% and the necessity of obtaining matching plasma samples, as in our previous procedure, could be circumvented. It is essential that each laboratory establish its own recovery factor because equipment and handling procedures will vary. A simple alternative to determining recovery factors would be to calibrate the instrument with DBS of known retinol concentration. Spots from whole blood with measured plasma retinol content could be prepared and stored for at least 1 wk. When extracted in the same way as the unknown samples, they can be used directly as calibrants to account for extraction efficiency and recovery during the first week of storage.

One criticism of previous DBS methods has been the emphasis on refrigerating the samples. This precaution was included because we were unsure of the stability of retinol under various conditions for longer time periods. For years, it was not expected that retinol, being labile to oxidation, would remain stable in DBS and especially not at room temperature. However, when retinol is bound to retinol binding protein, it is greatly protected from degradation (8 ). This is quite evident from the stability of retinol in DBS in this study. We observed that retinol in DBS is stable for 3 mo and longer when samples are stored at 23°C in the dark. Still it is prudent to maintain the samples at cooler temperatures, if possible. Of the storage conditions tested, we only found that the storage of inadequately dried DBS in a Ziploc bag without desiccant was detrimental to DBS retinol analysis. Storage under humid conditions has been previously observed to cause losses in DBS retinol. (Yinfa Ma, Truman State University, Kirksville, MO, personal communication)

The new procedure was applied to the analysis of retinol in DBS from India, Indonesia and Nicaragua. A subset of matching plasma samples was available from each location, which allowed us to calculate the correlation between DBS retinol and plasma retinol. In addition, the mean recoveries of retinol from this variety of DBS were estimated. The correlations between plasma retinol and DBS retinol were excellent with r² values > 0.95 and the recovery was fairly constant (75–84%) in all samples. Therefore, this method is a reliable means to measure serum retinol in large studies where plasma or serum samples are not readily obtained.

	FOOTNOTES

 
¹ Presented in part at the IVACG meeting, February 2001, Hanoi [Craft, N. E., Brindle, L. K. & Li, Y. Can dried blood spot retinol be used to assess vitamin A status?] and the ICN meeting, August 2001, Vienna [Craft, N. E., Brindle, L. K. & Li, Y. Using dried blood spot retinol to assess vitamin A deficiency]. Ann. Nutr. Metab. 45(suppl. 1): (abs. 4.03.011).

² Supported by grants from Institut Danone für Ernährung e.V., München/Germany; Micronutrient Operational Strategies and Technologies, Vienna, VA; Task Force Sight and Life, Basel, Switzerland; The Micronutrient Initiative, Ottawa, Canada; United Nations Children’s Fund, New York, New York; and the World Health Organization, Geneva, Switzerland.

⁴ Abbreviations used: BHT, butylated hydroxy toluene; DBS, dried blood spots; NIST, National Institute of Standards and Technology; VAD, vitamin A deficiency.

Manuscript received 11 June 2001. Revision accepted 24 October 2001.

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS Calculation of plasma volume... RESULTS DISCUSSION LITERATURE CITED

 

1. Underwood, B. & Arthur, P. (1996) The contribution of vitamin A to public health. FASEB J 10:1040-1048.[Abstract]

2. Sommer, A. (1998) Moving from science to public health programs: lessons from vitamin A. Am. J. Clin. Nutr. 68(2 suppl.):513S-516S.[Abstract]

3. World Health Organization (1994) Indicators for Assessing Vitamin A Deficiency and Their Application in Monitoring and Evaluating Intervention Programs: Report of a Joint WHO/UNICEF Consultation 1994 World Health Organization Geneva, Switzerland. .

4. National Committee for Clinical Laboratory Standards (1997) Blood collection on filter paper for neonatal screening programs: approved standard. 3rd ed 17:LA4-A3.

5. Craft, N. E., Haitema, T., Brindle, L. K., Yamini, S., Humphrey, J. H. & West, K. P. (2000) Retinol analysis in dried blood spots by HPLC. J. Nutr. 130:882-885.[Abstract/Free Full Text]

6. Craft, N. E., Bulux, J., Valdez, C., Li, Y. & Solomons, N. W. (2000) Retinol concentrations in capillary dried blood spots from healthy volunteers: method validation Am. J. Clin. Nutr. 72:450-454.

7. Scientific Tables Geigy (1979) Hematology 2 Ciba Geiga AG Basel, Switzerland. .

8. Oliver, R. W., Kafwembe, E. M. & Mwandu, D. (1993) Stability of vitamin A circulating complex in spots of dried serum samples absorbed onto filter paper. Clin. Chem. 39:1744-1745.[Medline]

9. Shi, H., Ma, Y., Humphrey, J. H. & Craft, N. E. (1995) Determination of vitamin A in dried human blood spots by high-performance capillary electrophoresis with laser-excited fluorescence detection. J. Chromatogr. B Biomed. Appl. 665:89-96.[Medline]

10. Cook, J. D, Flowers, C. H. & Skikne, B. S. (1998) An assessment of dried blood-spot technology for identifying iron deficiency. Blood 92:1807-1813.[Abstract/Free Full Text]

This article has been cited by other articles:

				M. Fujita, E. Brindle, J. Shofer, P. Ndemwa, Y. Kombe, B. Shell-Duncan, and K. A. O'Connor Retinol-Binding Protein Stability in Dried Blood Spots Clin. Chem., November 1, 2007; 53(11): 1972 - 1975. [Abstract] [Full Text] [PDF]

				L. G. Payne, K. G. Koski, E. Ortega-Barria, and M. E. Scott Benefit of Vitamin A Supplementation on Ascaris Reinfection Is Less Evident in Stunted Children J. Nutr., June 1, 2007; 137(6): 1455 - 1459. [Abstract] [Full Text] [PDF]

				J. W. Low, M. Arimond, N. Osman, B. Cunguara, F. Zano, and D. Tschirley A Food-Based Approach Introducing Orange-Fleshed Sweet Potatoes Increased Vitamin A Intake and Serum Retinol Concentrations in Young Children in Rural Mozambique J. Nutr., May 1, 2007; 137(5): 1320 - 1327. [Abstract] [Full Text] [PDF]

				T. W. McDade, J. Burhop, and J. Dohnal High-Sensitivity Enzyme Immunoassay for C-Reactive Protein in Dried Blood Spots Clin. Chem., March 1, 2004; 50(3): 652 - 654. [Full Text] [PDF]

				B. Shell-Duncan and T. McDade Use of Combined Measures from Capillary Blood to Assess Iron Deficiency in Rural Kenyan Children J. Nutr., February 1, 2004; 134(2): 384 - 387. [Abstract] [Full Text] [PDF]

				T. W. McDade and B. Shell-Duncan Whole Blood Collected on Filter Paper Provides a Minimally Invasive Method for Assessing Human Transferrin Receptor Level J. Nutr., December 1, 2002; 132(12): 3760 - 3763. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Erhardt, J. G. Articles by Biesalski, H. K. Search for Related Content PubMed PubMed Citation Articles by Erhardt, J. G. Articles by Biesalski, H. K.