1398 Am J C/in Nutr 1997:66:1398-405. Printed in USA. © 1997 American Society for Clinical Nutrition
Field-study screening of blood folate concentrations:specimen stability and finger-stick sampling1�
Sean D 0 ‘Broin, Brian P Kelleher, Ann Davoren, and Elaine W Gunter
ABSTRACT We describe optimized procedures for field
studies of blood folate concentrations by using finger-stick blood
sampling and include relevant studies on blood folate stability. We
introduce whole-blood folate adjustment using sample hemoglobin
(folate/hemoglobin, nmollg) as a novel and practical tool yielding
accurate and precise results when blood volume or dilution is
unknown. Red cell folate concentrations (nmollL) of 1 1 887
Americans correlated well with hemoglobin-corrected whole-
blood folate concentrations (r� = 0.993; red cell folate = 0.347 X
hemoglobin folate + I nmol/L), which supports the approach of
using the mean cell hemoglobin concentration (g/L) to interconvert
red cell and hemoglobin folate data. Folate concentrations in
capillary (finger stick) and venous blood samples from 28 normal
donors were similar (P > 0.87), correlating closely (r = 0.98, P <
0.001). Whole-blood samples (collected into K2-EDTA-contain-
ing evacuated tubes) in field studies are best stored intact at 4 #{176}C
until they can be processed and frozen (-20 #{176}C).Specific knowl-
edge of blood folate stability is essential in planning and designing
field studies. Am J Gum Nutr l997;66: I 398-405.
KEY WORDS Red cell folate, hemoglobin folate, driedblood spot, specimen stability, finger-stick blood sampling,
field studies
INTRODUCTION
Recent focus on folate status as a serious clinical and public
health issue has led to an increased demand for nutritionalsurveys of blood folate concentrations. Diminished folate sta-
tus may be associated with cervical (1) and colonic (2) carci-
nogenesis and is a primary determinant for homocysteinemia,
which may be an independent risk factor for atherosclerotic
vascular disease (3). The relations of these new clinical corre-
lates with folate status have yet to be evaluated fully. Because
data exist (4) that confirm the role of folate in the prevention ofspina bifida and anencephaly, extensive clinical and epidemi-
ologic studies are needed to support the development of effec-
tive primary prevention strategies for these birth defects (5).
To control costs in nutritional surveys, it is desirable tosimplify both blood collection and assay procedures when
possible. Red cell folate concentrations are unstable (6) andspecial care must be taken when deviating from standard pro-
tocols of collection and handling. For field studies in remote
areas, additional factors relating to climate, local facilities,
transportation, and storage can compromise blood folate sta-bility and must be evaluated in advance.
This study was initiated to simplify blood folate screening in
general and to develop effective logistics for conducting
screening in remote areas. The approach of analyzing folate
concentrations in capillary blood collected by finger-stick sam-
pling was selected because this method of sample collection
has proved to be both practical and economical compared with
venipuncture in field studies for other analytes (7).
MATERIALS AND METHODS
Materials
Type 903 filter paper cards (1 5 X 10 cm) for preparing dried
blood spots (DBS) were obtained from Schleicher and Schuell
(Keene, NH). Microvettes, cryovials, and microtiter plateswere from Sarstedt (Wexford, Ireland) and EDTA-anticoagu-
lated Microtainers (500 p.L) were from Becton Dickinson(Orangeburg, NY). EDTA-anticoagulated whole blood
(EDTA-WB) for complete blood counts (CBCs) was collected
into K2-EDTA-containing Vacutainers (Becton Dickinson) in
Ireland and into K3-EDTA elsewhere. Unistik 2 lancets were
from Owen Mumford Inc (Atlanta). Sodium ascorbate, ascor-
bic acid, Tween 80, Triton X-lOO, and sodium lauryl sulfate
were from Sigma Chemical Co (St Louis). A Decon FS-lOO
sonication bath was from Ultrasonics Ltd (Sussex, United
Kingdom). Serum femtin concentrations were determined with
the Abbott IMX system (Abbott Laboratories, Abbott Park,
IL). Serum vitamin B- 12 concentrations were measured by
microbiological assay on microtiter plates (8). Standard finger-
stick protocols were followed (9).
Folate assays
Red cell and serum folate concentrations were measuredwith the Bio-Rad Quantaphase II radioassay kit (Bio-Rad Lab-
I From the Department of Haematology, St James’s Hospital, Dublin;The Blood Transfusion Service Board, Dublin; and the Centers for DiseaseControl and Prevention, Atlanta.
2 Use of trade names is for the purpose of identification only and does
not constitute endorsement by St James’s Hospital, the US Public Health
Service, or the US Department of Health and Human Services.
3 Supported in part by the National Center for Environmental Health,
Centers for Disease Control and Prevention, Atlanta.
4 Address reprint requests to SD O’Broin, Department of Haematology,St James’s Hospital, Dublin 8, Ireland.
Received April 14, 1997.Accepted for publication July 17, 1997.
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
FIELD STUDIES OF BLOOD FOLATE CONCENTRATIONS I 399
oratories, Hercules, CA) and also with a microbiological assay
using Lactobacillus casei (NCIB 10463) on microtiter plates
(10). We compared the red cell folate concentrations in 1 18EDTA-WB samples using both assays. The samples repre-sented results of sequential diagnostic folate assays and in-
cluded never-thawed hemolysates from donors with poor folatestatus (n = 30), which had been stored at -70 #{176}C.Erythrocyte
5-methyltetrahydrofolate polyglutamates must be deconjugated
to the monoglutamate form before being assayed (1 1). Hydro-
lysis is achieved by lysing cells in ascorbic acid at a pH suitable
for the activity of endogenous plasma conjugase enzyme (‘y-glutamyl hydrolase, EC 3.4.12.10). Microbiological assay of
lysates of EDTA-WB, prepared by diluting the EDTA-WB 1:9
with 10 g ascorbic acid/L (eg, 100 p.L EDTA-WB + 900 �tL
of 10 g ascorbic acidlL), mixing, and then incubating theresulting hemolysate at room temperature for 30 mm, was used
as the 100% control whole-blood folate concentrations
throughout the study. In some stability studies low-folate se-
rum from folate-deficient patients was incubated with hemo-lysates as an additional conjugase source to establish endoge-
nous plasma conjugase sufficiency. In these cases, a volume of
low-folate serum was added that was equivalent to the endog-enous plasma volumes of the hemolysates.
Measurement of red cell indexes
CBCs were performed electronically on a Coulter STKScalibrated by using SC hematology reference controls (Coulter
Diagnostics, Hialeah, FL). Hemoglobin concentrations of
EDTA-WB samples from patient donors were measured by
using the reference cyanmethemoglobin method (12), and ly-
sates of these samples prepared as described were then used as
hemoglobin controls in later assays. Hemoglobin concentra-
tions in EDTA-WB in ascorbic acid (as prepared for red blood
cell folate determination) were quantified spectrophotometri-
cally (535 nm) after 1:20 dilution with a solution of sodium
lauryl sulfate (2.42 mmol/L) and Triton X- 100 (1 mLJL) in
0.03 mol sodium phosphate bufferlL (pH 7.3) mixed, and
allowed to stand for 3 mm. Hemoglobin concentrations weremeasured directly after precalibrating the spectrophotometerwith lysates of the hemoglobin control samples (n = 3).
Whole-blood hemoglobin samples (n = 83) measured with the
Coulter STKS correlated well with hemoglobin measured incorresponding hemolysates in 10 g ascorbic acidlL (r� = 0.97,
y = 1.03x - 0.29) and there was no significant difference
between the results (P > 0.7). Also, there was no significant
difference between hemoglobin concentrations in WB-EDTAsamples (n = 10) diluted 1:20 in water and those in WB-EDTAsamples diluted in 2.5, 5.0, 10, or 20 g ascorbic acid/L (P >
0.8). Sodium lauryl sulfate is used widely for conventionalhemoglobin measurements because of its comparatively lowtoxicity and because excellent comparisons have been obtainedwith the reference cyanmethemoglobin method (13, 14).
Dried blood spots
Blood specimens were stored as DBS on filter paper cards
(15 X 10 cm), each with 15 preprinted 1.25-cm circles. DBSwere prepared by pipetting SO p.L EDTA-WB onto the circles
on the cards, then placing the cards horizontally on a rack andallowing them to air-dry overnight at ambient temperature.Recovery studies indicated that > 90% of the original folate
concentration and > 96% of the original hemoglobin concen-tration were eluted from the spots prepared in this manner, after
the entire spot had been cut out with scissors and sonicated in
10 g ascorbic acidlL containing 1% (by vol) Tween 80 for 60mm. Some filter paper was preimpregnated with ascorbic acid
as an antioxidant in an effort to improve the stability of folate
in the DBS. This was done by immersing the paper card flat in10 mL 10-20 g ascorbic acid/L for 60 s, followed by air-dryingthe card horizontally on a rack.
Blood folate correction
We evaluated two approaches for adjusting whole-bloodfolate concentrations: the use of hemoglobin and the use of
traditional hematocrit correction. The close relation that exists
between these hematologic indexes is expressed as the mean
corpuscular (cell) hemoglobin concentration (MCHC) or he-
moglobin/hematocrit. The MCHC, as calculated by automated
analyzers as part of a CBC, has shown a remarkable constancy
across a range of ages in cord blood and in children and adults
(15, 16) of different ethnic groups, both males and females
(17).
Hematologic data, including serum folate and red cell folateconcentrations measured with the Bio-Rad radioimmunoassay
as well as CBCs from a nutritional survey of 1 1 887 healthyAmericans, were examined (EW Gunter, unpublished observa-
tions, 1994). These data were collected by using methods of
analyses described previously (I 8). Red cell folate concentra-
tions were derived from whole-blood folate and serum folate
concentrations by using the traditional hematocrit correction of
Hoffbrand et al ( 1 1) as follows:
Red cell folate (nmolIL)
- whole-blood folate - [serum folate (1 - hematocrit)] (1)
- hematocrit
Hemoglobin folate concentrations were determined by divid-
ing the whole-blood folate concentration by the concentration
of hemoglobin in the sample:
Hemoglobin folate (nmol/g)
- whole-blood folate_(nmollL) (2)
- hemoglobin (g/L)
We also analyzed the results of sequential diagnostic CBCrequests (n 2049) and nutritional screen requests (n = 868)
in the Department of Haematology of St James’s Hospital to
determine the constancy of the MCHC ratio in such samples.CBCs on EDTA-WB samples were estimated by using theCoulter STKS, and serum and red cell folate concentrations bymicrobiological assay.
Red cell and hemoglobin folate concentrations compared
Blood samples for red cell folate assay (n = 187) wereprepared by accurate volumetric dilution of fresh EDTA-WB
as described above. The same blood samples (n = 187) werediluted nonvolurnetrically by an independent laboratory tech-
nician who was supplied with tubes containing variable vol-umes of 10 g ascorbic acid/L and instructed to prepare hemo-
lysates by adding a few drops of blood to each tube. The
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
1400 O’BROIN ET AL
dilution factor of these blood lysates was thus unknown and
variable. Lysate folate concentrations were assayed by micro-
biological assay as described previously and hemoglobin con-
centrations determined from the lysates described previously
were used to correct the results of nonvolumetrically preparedsamples. Red cell folate concentrations of conventionally di-
luted blood samples were calculated from the hematocrit. The
folate concentrations of nonvolumetrically prepared lysates
were expressed as the folate concentration per unit of hemo-
globin (folate/hemoglobin, pmol/g); a dilution factor thus be-
comes irrelevant because it is common to both the numerator
and denominator.
Blood folate stability
Preliminary blood folate stability screens helped to approx-
imate rates of loss of microbiological assay activity and to
identify conditions suitable for further study. One such study
was made by using DBS stored at temperatures of -20, 8, 22,and 37 #{176}C.DBS were prepared by pipetting S0-pL aliquotsfrom three fresh EDTA-WB samples onto both plain filter
paper and paper that had been pretreated with 10 or 20 gascorbic acidlL as described. Blood spots were air-dried over-
night at ambient temperature and then stored in plain polyeth-
ylene bags without desiccant on day 1 . On subsequent days, up
to the completion of the study, DBS were removed fromstorage and placed at - 20 #{176}Cuntil extracted and assayed as
described previously.
Similar folate stability studies were made with EDTA-WB
samples (n = 14) stored as unlysed whole blood (K2-EDTA)
and as hemolysates in 10 g ascorbic acidlL at temperatures of
-20, 4, 22, and 37 #{176}C.Controls prepared as described were
stored at - 20 #{176}C.EDTA-WB samples were mixed and dis-
pensed in 50-pt aliquots into 1-mL cryovials and ascorbated
hemolysates of these samples were stored under the same
conditions. On days 1, 2, 4, 5, and 7, an aliquot of whole blood
was removed from storage, diluted with 10 g ascorbic acidlL
(1 : 10), mixed thoroughly, incubated at room temperature for 30mm, and stored at -20 #{176}Cuntil assayed. Blood lysates were
placed at - 20 #{176}Cat the same times and the folate concentra-
tions of all lysates and controls were measured in quadruplicate
in a single microbiological assay run.
Freeze-thaw stability of EDTA-WB folate was studied. In atypical experiment, control hemolysates were prepared and
l00-�L aliquots from 10 fresh EDTA-WB samples were fro-
zen overnight at -20 #{176}C,placed at room temperature (20 #{176}C),
and diluted by adding 10 g ascorbic acidlL immediately and
after 1, 6, and 24 h. After dilution, each lysate was mixedextensively by vortex, incubated at 20 #{176}Cfor 60 mm, and
frozen at - 20 #{176}C.Finally, all aliquots were thawed and the
folate concentration was measured by microbiological assay.
The stability of whole-blood folate at -20 #{176}Cwas observedas a between-assay reproducibility study. Aliquots (200 pL) ofEDTA-WB from three donors were frozen at -20 #{176}C.They
were removed, extracted as above, and assayed for folateweekly over 10 wk.
Factors influencing plasma conjugase activity
Stability
A study of EDTA-WB folate recovery included an evalua-
tion of plasma conjugase activity. Aliquots from six
EDTA-WB samples in cryovials were stored in batches at both4 and 22 #{176}C.On days 1, 4, 5, 6, and 7, an aliquot from eachbatch was stored at -20 #{176}C.Finally, all aliquots were diluted
1: 10 with 10 g ascorbic acidlL, incubated at 20 #{176}Cfor 30 mm,
and assayed. A portion of each aliquot was also incubated with
supplemental fresh low-folate plasma, as described previously.
Ascorbic acid concentration
Blood samples (n = 9) were diluted 10-fold in solutions of
20, 15, 12.5, 10, 7.5 and 5.0 g ascorbic acidlL. The dilutedsamples were incubated at 20 #{176}Cfor 30 mm, assayed for folate
concentrations, and the relative folate recoveries comparedwith those of the hemolysate in 10 g ascorbic acid/L. Blood
samples diluted in 10 or 5.0 g ascorbic acid/L containing 1%Tween 80 or 0.1% Triton X-100 were also compared with
controls.
Dilution factor
The relation between the degree of initial dilution of wholeblood in ascorbic acid and blood folate recovery was also
studied. Fresh blood samples (n 10) were diluted 1 : 10, 1:20,1:40, and 1:50 in 10 g ascorbic acid/L, then each dilution was
divided and supplementary fresh plasma conjugase was added
to one portion as a control before the samples were mixed and
incubated at 20 #{176}Cfor 30 mm. In a separate experiment, therelative folate recoveries of whole blood (n = 10) diluted 1:10,1:20, 1:50, and 1:100 in both 10 and 5.0 g ascorbic acid/L
containing 0.1% Triton X-100 were determined.
Finger-stick and venous blood folate concentrations
compared
Venous EDTA-WB samples were obtained from 28 labora-tory staff along with concurrent capillary blood finger-sticksamples collected into EDTA-containing cryovials after eachdonor’s hand was prewarmed as recommended (9). Lysates ofboth sets of blood samples were prepared by standard dilution
in 10 g ascorbic acidlL and assayed for folate. Finger-sticksamples in EDTA-containing cryovials were also obtained
from blood bank donors (n = 45). These finger-stick sampleswere taken from unwarmed finger-stick sites by donor atten-dants secondary to a primary sampling from the identical site
for hemoglobin screening. Blood was washed nonvolumetri-cally from these cryovials with a solution of 5.0 g ascorbicacid/L containing 0.1% Triton X-100, and venous control he-molysates were diluted 1 :20 in the same solution for compar-
ison. Similar fmger-stick samples from 56 blood bank donorswere collected directly into 1 mL of sterile 5.0 g ascorbic
acid/L and 0.1% Triton X-100 in cryovials. Donor attendantswere instructed to add a drop of blood directly from thefinger-stick site into the ascorbic acid, cap the vial, and mix thesolution immediately and thoroughly by inversion six times.Venous control sample lysates were prepared as described and
all samples were assayed together for folate either on the sameday, or else the samples were frozen at -20 #{176}Cwithin 3 h ofsampling and assayed later.
Statistical analysis
Results were compared by using the t test for paired obser-vations and by simple-regression analysis and are reported as
means ± SDs. The statistical package STATGRAPHICS (ver-
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
7000
8000
0.
0
I:0
r = 0.99
y=3.06x+ 118.4n = 187
2000ire
Red cell folate (nmoVL)ire
�2000
0
E
F500c8to(5� moo
0
0
�51xD
. 8,.#{149}
r = 0.84y = 1.8x + 107.8n = 218
0 200 400 800 800 000 200
Bio-Rad Quantaphase II (nmol/L)
FIGURE 2. Comparison of red cell folate concentrations estimated by
using a microbiological assay and by radioassay with Bio-Rad’s Quant-
aphase II (Bio-Rad Laboratories, Hercules, CA).
FIELD STUDIES OF BLOOD FOLATE CONCENTRATIONS 1401
TABLE 1
Comparison of red cell folate and hemoglobin folate concentrations in
healthy US subjects’
Males Females
Red cell folate (nmollL) 608.6 ± 278.9 [6045] 620.2 ± 313.2 [61 191
Hemoglobin folate 1744.8 ± 800 [5913] 1786. 1 ± 899.6 [59741(pmollg)
, I ± SD; n in brackets. Folate concentrations were assayed by using
Bio-Rad Quantaphase II; Bio-Rad Laboratories, Hercules, CA.
sion 1 .2; Statistical Graphics Corp. Rockville, MD) was used
for the analyses.
RESULTS
Distribution of MCHC in normal and disease states
In almost 12 000 healthy Americans, red cell folate andhemoglobin folate concentrations correlated well (r� 0.993,
P � 0.001; n = 11887) and were similarly distributed acrossthe full range of values (Table 1), thus reflecting the constancy
of the MCHC. This supports the use of hemoglobin for cor-
recting data and of MCHC for interconverting blood folatedata. The regression equations were as follows:
Red cell folate = 0.347 hemoglobin folate + 1.0 nmolIL (3)
Hemoglobin folate = 2.86 red cell folate + 10 pmollg (4)
Data from Ireland on 2049 sequential, diagnostic, full bloodcount requests supported this finding (MCHC = 343 ± 8 g/L)
but slightly lower MCHCs were noted in patients with lowserum ferritin concentrations (< 20 �g/L). In these patients(n 2 16) MCHCs (320.6 ± 1 7 g/L) correlated poorly with
serum ferritin concentrations (r = 0.2, P = < 0.01) but thiscorrelation improved with the degree of iron deficiency
(Table 2).
In the Irish data set there was no relation between severity of
red cell folate deficiency and MCHC (33 1 .6 ± 1 1 .6 g/L) in 457sequential patients selected on the basis of having simulta-
neously low serum ferritin (< 5.7 nmol/L) and red cell folate(< 340 nmolJL) concentrations. Also, MCHCs (335.5 ± 1 1.7gIL) of 481 sequential patients with a low serum vitamin B-l2concentration (< 1 10 pmol/L) correlated negatively with theserum vitamin B- 12 concentration when patients with low
serum femtin concentrations were excluded (r = -0. 1 16, P =
< 0.05).
TABLE 2
Mean cell hemoglobin concentrations of patients with iron deficiency,
correlated with serum ferritin concentrations
Serum femtin I ± SD r
gIL< 20 �tg/L (n = 216) 321 ± 17 0.2
< 15 p�g/L (n = 188) 320 ± 17 0.235
< 10 �g/L (n = 145) 319 ± 17 0.324
< 5 ,.Lg/L (n = 82) 316 ± 19 0.346
< 3 ;ig/L (n = 23) 305 ± 19 -
FIGURE 1. Comparison of traditionally prepared red cell and hemo-
globin folate concentrations from nonvolumetrically prepared lysates esti-
mated by using a microbiological assay.
Folate assays
Volumetrically prepared red cell folate concentrations (n
187) correlated well (r� 0.98, P < 0.001) with theirequivalent hemoglobin folate concentrations from lysates pre-pared by using dilutions of unknown proportions (Figure 1).Random nonvolumetric dilution of lysates for hemoglobin fo-
late was confirmed by comparing the whole-blood folate con-centrations from both data sets. Red cell folate concentrations
(n 2 1 8) determined by microbiological assay (85 1 .6 ± 465nmol/L) were substantially higher than those obtained by using
the Bio-Rad Quantaphase II radioassay (417 ± 219 nmol/L;
Figure 2.
Blood folate stability
Storage temperature was identified as a major influence onstability in all studies. In initial screens of blood folate stability,
lysates in 10 g ascorbic acid/L incurred an overnight folate loss
of > 50% at 37 #{176}C,thus this condition was not included in
subsequent studies. Comparatively good stability was achieved
for folate concentrations in DBS on plain filter paper but thesewere progressively less stable on filter paper that was pre-
treated with 10 or 20 g ascorbic acidlL and with increasing
2500
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
1402 O’BROIN ET AL
acid.
-�. 120
0
� 100
SCO
0
11)
>
02
0)C
C
CO
E0)
0)
CO0
LI- 0
FIGURE 3. Effect of temperature on folate stability when blood sam-
ples (n = 14) were stored as anticoagulated (with EDTA) whole blood (A)and lysates (B) in ascorbic acid. I ± 1 SD. Values < 10% are expressed
as 10%.
temperature (Table 3). A folate recovery > 90% was obtainedfrom 10 DBS on plain paper after 4 d at 4 #{176}C(90.5 ± 9.5%).
The addition of supplemental plasma conjugase to extracts
from DBS as described did not improve folate recovery.In general, folate recovery was related to storage tempera-
ture, whether blood was stored as EDTA-WB (Figure 3A), oras lysates in 10 g ascorbic acid/L (Figure 3B), or as DBS on
filter paper (Table 3). Maximum folate stability at 4 and 22 #{176}C
was achieved with intact whole blood (n = 14), with 73%recovery at 22 #{176}Cafter 5 d. Folate stability was poorest withlysed blood specimens. Results varied considerably in individ-
TABLE 3
80
60
40
20
ual samples stored as hemolysates in 10 g ascorbic acid/L, with
recoveries ranging from < 10% to 100% after 4 d at 22 #{176}C
(Figure 3B).Frozen and thawed EDTA-WB samples (n = 20) had 12%
lower folate concentrations than controls, which was signifi-cant (t = 3.79, P < 0.005). Folate concentrations ofEDTA-WB samples frozen at -20 #{176}C(n = 10) diminishedrapidly at 20 #{176}Cafter being thawed, with recoveries of 90%,
77.8%, and 51.2% after 1, 6, and 24 h, respectively. However,the folate concentrations of three intact EDTA-WB samples
that had been frozen in aliquots at -20 #{176}Cremained stable,with between-assay CVs of 12.0%, 8.6%, and 12.6%, after 10consecutive weekly assays.
Factors influencing plasma conjugase activity
Folate recovery from six EDTA-WB samples after storagefor 5, 6, and 7 d was 91%, 90%, and 92%, respectively, at 4 #{176}C
and 79%, 72%, and 66%, respectively, at 22 #{176}C.Folate recov-
ery was not improved by incubating hemolysates with supple-
mental fresh plasma, confirming the retention of sufficient
endogenous conjugase enzyme activity after such storage.Whole-blood folate concentrations were not significantly dif-
ferent when ascorbic acid at concentrations of 5.0, 7.5, 10,12.5, and 15 gIL were used to lyse blood samples (n = 9);when blood samples (n 10) were diluted 1 : 10 (controls),1 :20, 1:40, or 1 :50 in ascorbic acid; or when supplementary
plasma conjugase was added to these lysates. The folate recov-eries from blood samples (n = 10) diluted 1:10, 1:20, 1:50, or1:100, both in 10 g ascorbic acid/L and in 5.0 g ascorbic acid/L
containing 0.1% Triton X-l00, were virtually identical and the
inclusion of 1 % Tween 80 in the ascorbic acid solutions did not
influence recovery.
Finger-stick and venous blood folate concentrationscompared
Hemoglobin folate concentrations of volumetrically pre-pared finger-stick (2060 ± 663 pmollg) and of venous (2087 ±
657 pmoL/g) EDTA-WB samples from 28 normal volunteers
correlated well (,2 0.98) and were not significantly different(t = 0.155, P > 0.87; Figure 4). Folate concentrations of
finger-stick samples taken as second samples from unwarmed
sites (2205 ± 680 pmol/g) and diluted nonvolumetrically also
correlated significantly (r� = 0.9, P < 0.001) with venouscontrols (2178 ± 672 pmol/g) and were not significantly dif-
ferent (t = 0.86, P > 0.39). Estimation of hemoglobin folate
Mean percentage recovery of folate remaining in dried blood spots when stored for 1 wk at three different temperatures on both plain and ascorbated
filter paper’
Plain paper Paper A Paper B
8 #{176}C 22 #{176}C 37 #{176}C 8 #{176}C 22 #{176}C 37 #{176}C8 #{176}C 22 #{176}C 37 #{176}C
%
Day 1 87 83 59 99 75 44 86 62 21
Day 4 85 7 1 56 79 47 18 60 39 11
Day 5 80 75 58 72 47 18 62 28 11
Day6 77 71 51 71 44 18 60 29 10
Day 7 81 71 57 76 44 16 64 32 13
‘ Results are from three separate blood samples compared with a control. Paper A was pretreated with 1% ascorbic acid and paper B with 2% ascorbic
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
4000
A
.
0) 3500
0
E.9’ 3000
0)
CO.2 2500
03
12000
sire
0)0)
�ire
r = 0.98
y = 0.98x + 73.5n = 28
moo ire 2000 2500 3000 3500 4000
Venous control sample folate (pmol/g)
B7000
0) 6000
0
E.9��ooo
03
CO.90)
53000
�20O0
03
�ire
.
r = 0.92y = 0.94x + 236n = 56
0 ire 2000 3000 4000 5000 6000
Venous control sample folate (pmoVg)7000
FIELD STUDIES OF BLOOD FOLATE CONCENTRATIONS 1403
FIGURE 4. Comparison by microbiological assay of folate concentra-
tions in finger-stick blood samples collected into K3-EDTA (A) or 5.0 g
ascorbic acidlL (B) and venous blood samples.
overcomes the problem of accurately diluting such very smallsamples (< 300 p.L). Initially, 80% of small samples (< 100p.L) taken into EDTA showed small clots on microscopic
examination; this problem was eliminated by ensuring ade-quate mixing by shaking and inverting at least six times. Thefolate concentrations of finger-stick samples collected directlyinto sterile ascorbic acid solution (Figure 4) also compared wellwith venous controls (r� = 0.85, P < 0.001), giving similar
results (P > 0.3).
DISCUSSION
These studies were undertaken to support population screen-
ing of blood folate status in conventional settings as well as inremote areas with limited facilities, to simplify blood collectionand reduce costs without unduly sacrificing analytic accuracyand precision. Folate concentrations are unstable and studies ofblood folate recoveries under different conditions were consid-
ered important because few of the available studies were rel-evant and some were contradictory. Endogenous folate plasmaconjugase stability was a concern but did not limit folaterecovery under any conditions studied here; an inhibition of itsactivity was noted previously in sera from uremic patients (19)
The temperature of blood storage had a major influence onthe recovery of folate assayed microbiologically from all spec-
imen types, including intact whole-blood samples (Figure 3A),whole blood lysates (Figure 3B), and DBS (Table 3). Also,
folate concentrations in all unfrozen blood samples were morestable in the intact red cell than as a hemolysate. The generally
decreased stability of blood folate concentrations as lysates orDBS in the presence of ascorbic acid was disconcerting. Alikely explanation is that ascorbic acid rapidly lyses the red
cells and that in its absence the intact cell affords greaterprotection to labile folate concentrations.
Storage as unfrozen intact EDTA-WB offered the best over-all folate stability, yielding recoveries of 86% on storage at
4 #{176}Cfor 7 d (Figure 3A) and 92% at 4 #{176}Cin a follow-up study(n 6). Previous studies at 4 #{176}Cnoted either no deterioration
in the folate concentrations of similar samples after storage for7-10 d (I 1) or for 14 d (20), or a 20% loss after 3 d when
shaken (6). Blood samples that cannot be processed immedi-ately should be stored in this manner at a temperature > 0 #{176}C(eg, 4 #{176}C),a temperature at which folate degradation is slow.
The relative instability and erratic recoveries of folate from
ascorbate-treated blood hemolysates (Figure 3B) confirm thefindings of previous studies (6, 1 1), despite one report to the
contrary (21). These problems limit the potential of preparing
such hemolysates in some field-study settings because the
samples need to be either assayed immediately or frozen at
-20 #{176}C(or -70 #{176}C).These hemolysate folate concentrations
will remain stable when frozen at -20 #{176}Cand have served wellas long-term quality control materials (22, 23).
Contrary to the findings of a previous study (1 1), we foundthat folate concentrations in EDTA-WB samples at - 20 #{176}C
remained stable for � 10 wk despite poor between-assay re-
producibility (CV: > 10%) relative to that of conventional
microbiological assay controls (22). Difficulties in resuspend-
ing these viscous frozen and thawed samples may account for
the sporadic errors that occurred; a systematic folate loss of� 12% was noted as a result of the freezing and thawing. Also,
when thawed, the folate concentrations of frozen EDTA-WBsamples degrade rapidly with almost a 50% loss of activityoccurring at 20 #{176}Cover 24 h; therefore, blood samples thatthaw during transportation should not be used.
Traditional whole-blood folate correction by using hemato-crit was introduced because red cells can contain SO times the
amount of folate that serum contains (24) and red cell volumesvary. The usefulness of correcting whole-blood folate concen-trations by using the sample hemoglobin became apparent early
in the present study; previous work had suggested the use of aderived hematocrit value to estimate red cell folate concentra-tions (25). Blood folate concentrations may be adjusted by
using the hemoglobin concentration because the MCHC ratio is
physiologically controlled within relatively narrow limits (26).
This relation has been well established both in clinical hema-tology (15, 16) and in the establishment of hematologic refer-
ence data for normal populations ( 17). Also, an internationalhematologic study of consecutive patients admitted to teachinghospitals in Wales, California, and Japan (n = 3500) noted nosignificant geographic difference in mean MCHC values (26).
The potential of using population mean MCHCs to manipulate
blood folate data was validated in our analysis of data fromalmost 12 000 free-living American subjects (Table 1) becausered cell folate concentrations and hemoglobin folate concen-
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
1404 O’BROIN ET AL
trations correlated well (r� = 0.993). Little variation in MCHCwas noted in an analysis of 2049 consecutive diagnostic CBC
requests: MCHC 343 ± 8.2 gIL, comparing well with mean
hematologic reference data (17) established for males (343 ±
12.5 g/L; n = 8449) and females (338 ± 12.9 g/L; n 8829)
aged 3-74 y. It is apparent that the CVs obtained for MCHCsin these hematologic screens (CV: < 5%) can be less than
those obtained for within-assay reproducibilities of many ana-
lytes. Falsely elevated MCHCs (> 360 g/L) occasionally arise
in our diagnostic CBC screens when the turbidity of hyperli-
pemic plasma causes falsely elevated hemoglobin concentra-tions (27) or when hematocrit is calculated incorrectly because
of the presence of cold agglutinins (28) and needs adjustment.Moderately altered MCHCs, which may be seen in iron defi-
ciency (Table 2), may result in small distortions in the inter-conversion of blood folate data, but this effect is difficult toquantitate. Studies of MCHC in disease are hampered by thetechnical difficulties associated with hematocrit measurement,which have been reported extensively (29-35), and by the
absence of an international standard for hematocrit (30). Nosuch technical shortcomings have been associated with the
estimation of hemoglobin, which is a simple colorimetric assay
whether measured manually or by automated methods. Inter-national reference preparations exist for hemoglobin assay
calibration (1 1), and blood folate data adjusted by using thesample hemoglobin will have less interlaboratory variation.
The minimum objective of these studies was to develop amethod for evaluating screening of compliance with folatesupplementation. A substantial lack of agreement exists cur-rently between the results of readily available conventional
serum and red cell folate assays (36) and is noted here in a
comparison of results between the radioassay and microbiolog-
ical assay of red cell folate concentrations (Figure 2). Because
the results of these analyses are method specific (36), we
considered that any method that was practical and optimized
for folate stability in the field would have potential for devel-
opment if it gave reproducible results in the laboratory. A
nonvolumetric folate screening assay with DBS has such po-tential; studies of folate stability with use of DBS are a pre-
requisite to further development and show some promise (Ta-
ble 3). A report on folate analysis in DBS is forthcoming.Simple finger-stick blood sampling emerged as being suit-
able even for conventional volumetric analysis of blood folate
concentrations (Figure 4A), and is useful in any setting. Accu-rate blood folate concentrations can also be obtained when the
blood volume or dilution factor is unknown (Figure 1), thus
allowing folate analysis of small finger-stick samples or those
collected directly into ascorbic acid solution to form hemoly-sates (Figure 4B); hematocrit measurements and dilution fac-
tors can now be ignored. This flexibility should be particularlyadvantageous under field conditions, where accurate dilution ofblood can be difficult. Red cell folate and hemoglobin folate
data become interconvertible later when applying either spe-
cific or population mean MCHCs. In certain studies, ideally, a
plasma or serum folate concentration is needed to calculate the
red cell folate concentration (37); however, this is not possiblewith very small or lysed finger-stick samples; these small
specimens may be suitable, however, for evaluating compli-ance in folate-intervention trials.
In field surveys, standard techniques often have to be mod-ified to suit local conditions and to meet financial resources. It
is important to identify the variables associated with these
changes and to compensate for them. For blood folate screens,
the suitability and economy of finger-stick peripheral bloodsampling coupled with the flexibility and stability of correction
by using sample hemoglobin should prove useful. However, an
awareness of the instability of blood folate concentrationsremains crucial in the design of field-study protocols. U
We thank Emer Lawlor and Joan 0’ Riordan of the Blood Transfusion
Service Board and the staff of the Centers for Disease Control and Pre-
vention (Atlanta) and St James’s Hospital (Dublin) for their help in
providing blood samples.
REFERENCES
1. Butterworth CE, Hatch KD, Soong SJ, et al. Oral folic acid supple-
mentation for cervical dysplasia: a clinical intervention trial. Am J
Obstet Gynecol 1992;166:803-9.2. Lashner B. Red blood cell folate is associated with the development of
dysplasia and cancer in ulcerative colitis. J Cancer Res Clin Biol
1993;l 19:549-54.3. Green R, Jacobsen DW. Clinical implications of hyperhomocysteine-
mia. In: Bailey LB. ed. Folate in health and disease. New York: Marcel
Dekker, Inc. 1995:75-122.
4. MRC Vitamin Research Group. Prevention of neural tube defects:
results of the Medical Research Council Vitamin Study. Lancet
l99l;338: 13 1-7.
5. Mulinare J, Cordero JF, Erickson JD, Berry Ri. Periconceptional use
of multivitamins and the occurrence of neural tube defects. JAMAl988;260:3 141-5.
6. O’Broin SD, Temperley U, Scott JM. Erythrocyte, plasma and serum
folate: specimen stability before microbiological assay. Clm Chem
l980;26:522-4.
7. Slazyk WE, Phillips DL, Therell BL, Hannon WH. Effect of lot-to-lotvariability in filter paper on the quantification of thyroxin, thyrotropin,
and phenylalanine in dried-blood specimens. Clin Chem 1988;
34:53-8.
8. Kelleher BP, O’Broin SD. Microbiological assay for vitamin B12
performed in 96-well microtitre plates. J Clin Pathol
1991 ;44:592-5.
9. National Committee for Clinical Laboratory Standards. Procedures for
the collection of diagnostic blood specimens by skin puncture: ap-proved standard. Villanova, PA: NCCLS, 1991. (NCCLS document
H4-A3.)10. O’Broin 5, Kelleher B. Microbiological assay on microtitre plates of
folate in serum and red cells. J Clin Pathol l992;45:344-7.
11. Hoffbrand AV, Newcombe BF, Mollin DL. Method of assay of red
cell folate activity and the value of the assay as a test for folate
deficiency. J Clin Pathol l966;19: 17-28.
12. National Committee for Clinical Laboratory Standards. Approved
standard. Reference procedure for the quantitative determination of
hemoglobin in blood. Villanova, PA: NCCLS, 1991 . (NCCLS ap-
proved standard HIS-A.)
13. Oshiro I, Takenaka T, Maeda J. New method for hemoglobin deter-mination by using sodium lauryl sulphate (SLS). Clin Biochem
l982;74:83-8.
14. Karsan A, Maclaren I, Conn D, Wadsworth L. An evaluation of
hemoglobin determination using sodium lauryl sulfate. Am J ClinPathol 1993;l00:l23-6.
15. Abelson HT. Approach to the child with anemia. In: Hoffman R, Benz
El, Shawl SJ, Furie B, Coen RI, eds. Hematology: basic principles andpractice. New York: Churchill Livingstone, 1991:311-9.
16. Williams WJ, Schneider AS. Examination of the peripheral blood. In:
Williams WJ, Beutler E, Erslev Al, Rundles RW, eds. Hematology.
2nd ed. New York: McGraw-Hill, 1977:10-25.
17. National Center for Health Statistics, Fulwood R, Johnson CL, et al.
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021
FIELD STUDIES OF BLOOD FOLATE CONCENTRATIONS 1405
Hematological and nutritional biochemistry reference data for persons
6 months-74 years of age: United States, 1976-80. Vital Health StatSer l982;232. [DH}IS publication no. (PHS) 83-1682.1
18. Gunter EW, Lewis BG, Koncikowski SM. Laboratory methods usedfor the third National Health and Nutrition Examination Survey(NHANES ifi), 1988-1994. Atlanta: Centers for Disease Control and
Prevention, 1996.19. Livant El, Tamura T, Johnson KE, et al. Plasma folate conjugase
activities and folate concentrations in patients receiving hemodialysis.J Nutr Biochem l994;S:504-8.
20. Gilois CR, Stone J, Lai AP, Wierzbicki AS. Effect of haemolysate
preparation on measurement of red cell folate by a radioisotopic
method. J Clin Pathol 1990;43:160-2.21. Chanarin I. The megaloblastic anaemias. London: Blackwell Scientific
Publications, 1969.
22. Kelleher BP, O’Broin SD. Choice of materials for long-term quality
control of blood folate assays. Clin Chem 1996;42:652-4 (letter).23. Gunter EW, Twite DB. Improved materials for long-term quality-
control assessment of erythrocyte folate analysis. Clin Chem l990;36:
2139 (letter).
24. Chanann I. The megaloblastic anaemias. London: Blackwell Scientific
Publications, 1979.25. O’Connor G, Molloy A, Daly L, Scott JM. Deriving a useful packed
cell volume estimate from haemoglobin analysis. J Clin Pathol
l994;47:78-9.
26. Bull BS, Hay KL. Are red blood cell indices international? Arch Pathol
Lab Med 1985;109:604-6.
27. Kalache GR, Saflor MM, Hughes WG. The indirect measurement ofhemoglobin concentration in whole blood. Pathology
1991;23:1 15-7.
28. Schwartz SS, Berkman EM, Silberstein LE. The autoimmune hemo-
lytic anemias. In: Hoffman R, Benz El, Shattil SJ, Furie B, Coen HJ,
eds. Hematology, basic principles and practice. New York: Churchill
Livingstone, 1991:422-41.29. National Committee for Clinical Laboratory Standards. Procedure for
determining packed cell volume by the microhematocrit method. Vil-
lanova, PA: NCCLS, 1988. (NCCLS document H7-A2.)
30. Thom R. Automated red cell analysis. In: Cavill I, ed. Measuring
techniques in haematology and comparative assessment of haematol-
ogy analysers. Ballieres Clin Haematol 1990;3:837-5031. International Committee for Standardisation in Haematology Expert
Panel on Blood Sizing. Recommendations for reference method for
determination by centhfugation of packed cell volume of blood. J Clin
Pathol 1980;33: 1-2.
32. England JM, Walford D, Waters DAW. Re-assessment of the reliabil-
ity of the haematocrit. Br J Haematol 1972;23:247-56.
33. Guthrie DL, Pearson TC. PCV measurement in the management of
polycythaemic patients. Clin Lab Haematol 1982;4:257-65.
34. Patton WN, Cave RI, Harris RI. A study of changes in red cell volume
and haemoglobin concentration during phlebotomy induced iron defi-
ciency and iron repletion using the Technicon Hi. Clin Lab Haematoll991;13:l53-6l.
35. Paterakis GS, Laoutaris NP, Alexia SV, et al. The effect of red cell
shape on the measurement of red cell volume: a proposed method for
the comparative assessment of this effect among various haematology
analysers. Clin Lab Haematol l994;l6:235-45.
36. Gunter EW, Bowman BA, Caudill SP, Twite DB, Adams MJ,
Sampson El. Results of an international round robin for serum and
whole-blood folate. Clin Chem 1996;42: 1689-94.
37. Kelleher BP, O’Broin SD. High serum folate and the calculation of red
cell folate. Clin Lab Haematol 1995;17:203
Dow
nloaded from https://academ
ic.oup.com/ajcn/article/66/6/1398/4655975 by guest on 30 D
ecember 2021